Tech Trend Radar

How will technology impact insurance

in 2023?

Image: BullRun / stock.adobe.com

Tech Trend Radar 2023
Introduction

The Tech Trend Radar 2023 provides

information about technology-driven
trends that are relevant to insurance
companies in the global insurance
sector. Made by insurers for insurers,
the radar aims to sharpen awareness,
provoke discussion and initiate new
business opportunities that appeal to
all insurance clients.

It is a collaborative initiative by Munich

Re Business Technology and ERGO IT
Strategy. Together with Business Units
and Strategic Units within Munich Re
Group and external experts such as
Institute of Electronic Business e. V.,
Forrester and Gartner, future trends
have been gathered, aggregated and
rated in order to provide a
comprehensive view of technology
trends, their maturity and relevance
for Munich Re Group and the insurance
industry.

The Tech Trend Radar 2023 provides an

overview of innovative trends and
solutions in the global insurance market.

2
Tech Trend Radar 2023
Contents
Foreword 4 Connected
Healthy Human 10 Experience 37 Data & AI 63 Cyber & Crypto 105 Redefining Industries 137
How to read this report 7
Behavioural Analytics 15 Automotive IoT 42 AI Democratisation 70 Cyber Security 110 Robotic Revolution 142
Tech Trend Radar 2023 9
Digital Healthcare 17 Data Fabric 44 Digital Twin 72 Digital Identity 119
Cyber Security 113
Edge Computing
Adopt
Highlight 46 Machine-driven Decisions 74 Human Recognition 121
Regional pages 165 Natural-language Processing 77
Next Gen Insurance 180 Trustworthy AI 79

AI Medicine 19 Ambient Computing 48 Data Mesh 82 Deepfake Defense 123 Autonomous Things 144
Biomanufacturing 21 Extended Reality 50 Knowledge Graphs 85 Digital Currencies 125 Hydrogen Economy 146
Trial

Personalized Medicine 23 Large Language Models 87 Smart Contracts 127 New Energy at Scale 148
Population Health 25 Quantum Computing 89 Privacy-enhancing Tech 129 Smart Cities 150
Robotic Health 27
Human Enhancement 29 Brain- 52 Multimodal AI 92 Automated Compliance 131 Blue Economy 152
Vaccine Technologies 31 machine Interface Reinforcement Learning 95 Decentralised Data Economy 133 Business Resilience 154
Assess

Metaverse 54 Carbon Capture 156

Super Apps 56 In-Space Economy 158
Ubiquitous 58
Connectivity
Artificial General Intelligence 97
Hold

Our approach 6
Contribution 188
Imprint 189 Our Solutions 33 Our Solutions 60 Our Solutions 99 Our Solutions 135 Our Solutions 160

3
Tech Trend Radar 2023
Foreword

Message from the CEO

Recent advances, particularly in the field of artificial intelligence, Our experts have assessed the most important future
have the potential to revolutionise the way we work and interact trends in these areas to provide a comprehensive
with one another. Keeping up with the latest trends in technology overview of their maturity and relevance to the
is vital for companies in almost every sector to stay agile and insurance industry.
meet evolving client needs.
We invite you to join us for a deep dive into these topics
To be competitive, though, it takes more than that. It calls for the and the trends that will shape our business of the future.
ability to adapt to this ongoing change in a way that is
Sincerely yours,
sustainable and profitable for your business, to apply expert
knowledge in a targeted manner and to act with foresight.
Innovations such as AI, natural-language processing or quantum
computing can help insurers automate processes, analyse data
more effectively, and provide personalised services to clients.
Our Munich Re Tech Trend Radar, now in its tenth year, is Joachim Wenning
designed to help you keep track of the latest developments. We
Chair of the Board of Management
have identified – and offer appropriate solutions for – five trend
fields featuring game changing technological progress: “Healthy
Munich Re
Human”, “Connected Experience”, “Data and AI”, “Cyber and
Crypto”, and “Redefining Industries”.

4
Tech Trend Radar 2023
Foreword

Optimise core capabilities and implement cultural changes

Martin Thormählen
Our Tech Trend Radar team was in full swing to analyse Optimise your core capabilities and stop tinkering with AI Chief Technology Officer
the selected new trends for our 10th anniversary edition, New digital distribution partners, brokers, agents and TPAs continue Munich Re
when ChatGPT from OpenAI arrived in November 2022 and to invest into backend automation and API driven business models.
now the world has a new ‘cambrium explosion’ to explore. Your underlying architecture has to support this now! And it mthormaehlen@munichre.com
Before we briefly touch on the advances in AI let us first explain what improves your overall as well.
you can expect from this year’s 10th anniversary edition: has been taken to a new level by Sam
Culture of collaboration at scale: Our global team of authors has Altman’s OpenAI and their partnership with
grown significantly as we not only analyse and classify trends, but Microsoft. We expect significant advances in document under-
Munich Re and ERGO experts from all business and central units standing and processing soon by various companies – having the
continuously explore innovation potential and build solutions. You new AI regulations at the heart of everything we do in insurance.
might imagine the great pleasure to discuss the trends with the The news don’t stop here – there is so much more to explore:
experts – you will find the names and key insights of many of them
 Self-driving vehicles  , which touches on risks
in the new radar.
 Devices understand humans 
Regional aspects: We’ve heard you loud and clear in our webinars, on
conferences and one-on-ones with CEOs and we deliver: This year we  Innovations in Oceans 
have added regional pages to the Tech Trend Radar to even better  Life & Health  explore the updates in the
Daniel Grothues
explain regional differences. trend field Chief Technology Officer
Next generation insurance – Post-Millennials: Originated from a  Big updates on sustainability trends in the ERGO Technology &
discussion with Dutch insurers we have also asked our young talents field Service Management
to find out how Post-Millennials will buy insurance. All insurers
agreed that also the young generation will seek insurance, but the Culture of insurance organisation must adopt the change
daniel.grothues@itergo.com
question is if they will continue to choose established insurers or buy Unlike 10 years ago, we now see clearly that we have an abundance
coverage differently in the future. of technology choices for any given need. The leadership team is in
charge of adjusting the culture to embrace change that leverages the
opportunities while leaving the core values untouched.
Let’s develop our future – together

5
Tech Trend Radar 2023
Our approach

Step 1: Screening
Analysis of trend developments
Result
Compiling developments and new trends for 2023 with external analysts’ reports
and internal market know-how, following 3 Golden Rules:
1. Technologies that potentially disrupt the insurance industry. Total Trends
2. Technologies that potentially change business models.
3. Technologies that potentially have an influence on the value chain
and on internal and external processes within the insurance industry.

Step 2: Aggregation and trend selection

Definition of trends
Aggregating data from screening processes and defining the most
relevant trends categorised in five primary fields – and screening for Selected Trends
“outdated” trends.
Further drill-down and validation with market data and identification of
corresponding use cases.
Disclaimer: References to companies do not constitute or imply
endorsement of any company or organisation.

5 Trend Fields with

Step 3: Evaluation 48 Trends
Assessment of impact and relevance (11 new trends)
Classifying trends according to their level of relevance for ERGO and Munich Re. (7 adopted or transformed trends)
Analysing trend impact along the insurance value chain and deriving opportunities and risks.
6
Tech Trend Radar 2023
How to read this report
The report is structured along five trend fields.

Healthy Connected Data and AI

Human Experience

Cyber and Redefining

Crypto Industries

Trends are provided with a trend maturity.

Hold
Add it to your watch list

Assess
Adopt Trial Assess Hold
Consider what it may
mean for your business

Explore the interactive radar Trial

First initiatives should be underway
The interactive radar may be filtered by line of business to quickly in the most affected business areas
find what is most relevant for you. Each trend comes with a short
Adopt
description, opportunities and risks, as well as selected cross-industry Start taking full advantage
use cases. Most importantly, the report provides an assessment of a of this technology
trend’s maturity.

7
Tech Trend Radar 2023
This year’s changes

New trends 2023 Adopted or transformed trends

 Ambient Computing  Advanced Batteries (adopted)
 Automotive IoT  AgTech (adopted)
 Blue Economy  Computer Vision (adopted)
 Business Resilience  Digital Assistants (adopted)
 Carbon Capture  Enhanced Reality (transformed into
 Data Mesh Extended Reality)
 Hydrogen Economy (re-added)  Trusted AI (transformed into Trustworthy AI)
 Large Language Models  Cellular Smart Grid (transformed into
New Energy at Scale)
 New Energy at Scale
 Super Apps
 Ubiquitous Connectivity

Adopt Trial Assess Hold

You would like to dig deeper on chosen trends? Explore the interactive version?
Our aim is to promote innovative initiatives around each tech trend with our clients Access the augmented reality experience using the
and develop the best solution for the challenges of tomorrow. Please feel free to reach Snoopstar app or access the interactive website via
out to us with any questions and to your responsible client managers for the full munichre.com/ttr2023
report, which additionally features three use cases per trend.

8
Tech Trend Radar
2023
Artificial General Intelligence

Multimodal AI

Reinforcement Learning

Ubiquitous Connectivity

+
Decentralised Data Economy
Quantum Computing
Metaverse
Large Language Models

+
Automated Compliance
Super Apps
+
Data Mesh

+
Extended Reality Smart Contracts
Knowledge Graphs
Brain-Machine Interface

Digital Currencies
Ambient Computing

+
AI Democratisation
Privacy-Enhancing Tech

Machine-Driven Decisions
Cyber Security Blue Economy

+
Trustworthy AI
Vaccine Technologies Edge Computing Deepfake Defense
In-Space Economy
Digital Twin
Automotive IoT
+

Robotic Health Smart Cities

Human Recognition
Data Fabric NLP Business Resilience

+
Biomanufacturing
Personalized Medicine Digital Identity Autonomous Things

Population Health Carbon Capture

+
New Energy at Scale

+
Digital Healthcare
AI Medicine

Human Enhancement Robotic Revolution Hydrogen Economy

+
Behavioural Analytics

Hold Assess Trial Adopt Contact Martin Thormählen Daniel Grothues

Watch list Evaluation Initiatives in Start initiatives new in 2023 Chief Technology Officer Chief Technology Officer
needed affected units in your unit mthormaehlen@munichre.com daniel.grothues@itergo.com
9
Healthy Human
Technology impacts in 2023

The global
Around
telemedicine market
US$ 809.2bn is expected to reach

US$ 123bn
is the projected digital
The global health-
health market size care biometrics markets
by 2030. is predicted to grow to
by 2030.
FDA
US$ 32.9m
considers by 2030. The global
lab-grown meat The global medical non-oncology precision
made from cultured chicken robotics market is medicine market is forecast
expected to grow to
cells safe for human to more than triple to
US$ 44.45bn

US$ 130bn
consumption in late
by 2030.
2022.
by 2030.
All trends 2023
Healthy Human
Healthy Human as a trend field considers all technologies that are designed to make our lives healthier. Vaccine technologies and drug development is an exemplary state-of-the-art medical research field which is
We are observing the development of digital healthcare and behavioural analytics which enable better aimed at evolving vaccines for therapeutic purposes. The Healthy Human trends in 2023 primarily address
coordination between different service providers and health status monitors. insurers that offer life and health insurance products. It is important for them to understand that, with new
Now that artificial intelligence in medicine (generic and personalised) and population health is being technologies in digital health, telemedicine and even the field of biotech, demand for new insurance solutions
applied to deliver care in a proactive manner, we are seeing increasingly more new diagnosis and will not only increase but key functions, like underwriting and risk management, can be enhanced with new
treatment approaches being developed for diseases. We are also observing development in Robotic data streams and insights (e.g., with population health data).
Health, which are enabling more subtle movements during complex operations or in therapy.

Adopt – Start initiatives in your unit Trial – Initiatives in affected units Assess – Evaluation needed Hold – Watch list
Behavioural Analytics AI Medicine Human Enhancement
Behavioural analytics monitor, analyse, measure and Machine Learning and other AI allow a better and more Technology is used to improve or augment physical,
interpret people’s actions, intentions and characteristics objective data analysis and speed up medical processes. cognitive or psychological abilities, such as through brain
from users’ footprints across devices. machine interfaces, brain computer interfaces or
exo-skeletons.
Biomanufacturing
Digital Healthcare The industrial production of a (bio) product using
Advanced technologies enable completely new biological organisms and systems such as cells or Vaccine Technologies
approaches to health and wellbeing. microorganisms for various industries from pharma New methods to produce safer and more effective
to energy. vaccines in a faster, cost-effective way as they do not
require the use of live or attenuated viruses or bacteria.
Personalized Medicine
Increasing AI options and medical advancements allow
individual and personalised treatments with much higher
effectiveness.

Population Health
Population health refers to the health outcomes of a
group, based on data of many individuals in defined
populations such as nations, age groups, cultures.

Robotic Health
Cognitive AI in robotics looks promising for the use of
robots in care, rehabilitation and prosthetics, as well
new in 2023
as surgery.

12
Trend Evolution
Healthy Human

Evolution over time

Adopt Trial Assess Hold

Start initiatives in your unit Initiatives in affected units Evaluation needed Watch list

AI Medicine 2023 2022

Behavioural Analytics 2023 2022 2021 2020

Biomanufacturing 2023 2022

Digital Healthcare 2023 2022

Human Enhancement 2023 2022 2021 2020 2019

Personalized Medicine 2023 2022 2021

Population Health 2023 2022 2021

Robotic Health 2023 2022 2021 2020

Vaccine Technologies 2023 2022 2021

13
Impacts on the Value Chain
Healthy Human

Explanation Insurance Value Chain

The illustration represents the impact the trend
fields have upon on the insurance value chain. Product Sales & Risk Customer Asset
With the dark boxes reflecting areas of high Design & Pricing Distribution Underwriting Policy Admin Management Engagement Claims Management
impact, and light boxes low impact.

AI Medicine

Behavioural Analytics

Biomanufacturing

Digital Healthcare

Human Enhancement

Personalized Medicine

Population Health

Robotic Health

Vaccine Technologies

14
Behavioural Analytics
Adopt
Healthy Human

Behavioural analytics monitor, diabetes or high blood pressure which Opportunities

require not only medication but changes
analyse, measure and interpret  Health and lifestyle data can be used as  Fraud prevention by detecting health
to the lifestyle that is causing the disease.
people’s actions, intentions key information to enable enhanced risk risks such as smoking that were not
Subsequently, this behavioural data can
and characteristics from users’ assessment in comparison to static disclosed in risk questionnaires or by
support patients in better managing
questionnaire responses uncovering multiple claims for the
digital footprints across websites, their everyday life.
 Customer segmentation based on same condition
mobile apps, and digital products. When future health problems and behaviour, such as health and wellness  Usage of new customer engagement
While behavioural analytics are not diseases can be forecasted based on habits, products and services can be channels as data may show that a certain
anything necessarily new, their relevance behavioural data like physical inactivity, tailored accordingly type of life or health insurance
is continuing to grow in parallel with the excessive social media use, or non- is required
 Personal habits and health records can be
amount of data generated across multiple compliance with treatment plans,
used to more effectively and individually
data-sources, such as from social media behavioural analytics becomes even
price health insurance products
platforms. The same applies to the growing more appealing for insurers.
number of digital health applications and
wearables that are used worldwide.
Risks
Today, most modern smartphones or
smartwatches are already equipped with  Bias, either in the data or in the
different sensors, such as measuring the algorithm design, can result in
number of steps, blood pressure or blood unfair treatment of certain groups
oxygen levels. Concurrently the tools used of customers
for behavioural analytics have developed  Companies collecting and analysing
further, being able to aggregate, analyse behavioural user data are required to
and process more data along with be compliant with relevant laws and
contextual information. This is especially regulations and need to protect the
relevant for healthcare where behavioural privacy and personal data of their users
data such as blood pressure, step counting  Fake or incorrect behavioural data can
or meal tracking can be used for a more lead to wrong risk- or pricing
holistic health assessment and monitoring assumptions
of chronic diseases such as
15
Behavioural Analytics
Adopt
Healthy Human

Roadmap for Home The “U-Scan” is a new at-home urine analysis device. This cartridge-based platform is able to detect,
measure, and analyse multiple biomarkers contained in urine.
Behavioural Health Urine Depending on the cartridge utilised, the luteinizing hormone for menstrual cycle prediction as well as

Integration “Lab” ovulation windows can be determined or values like water balance, pH levels or ketones for judging
energy metabolism and vitamin C levels can be measured. It is intended to encourage and/or help
Source Source the user maintain a healthy lifestyle by being able to independently and regularly self-monitor
ASPE & HHS Withings themselves using the data monitored.
France SA The “U-Scan For Professionals” cartridge is an at-home solution to give healthcare professionals
The U.S. Department of Health & access to regular and accurate urinalysis data to easily monitor patients and health program
Human Services (HHS) launched a participants or to assist medical research, clinical trials or remote patient monitoring.
roadmap that aims to ensure the full
spectrum of behavioural health care
will be integrated into the health care,
social service, and early childhood
systems with the goal of ensuring
Key Insight
people having equitable access to
evidence-based, culturally Rising volume of behavioural health data and new data-sources
appropriate, and person-centered
allow for more in-depth analysis and insights
care. The roadmap also builds upon
US$ 13 million in federal funding to Insurers need to deal with the growing volume of aggregated
increase access to behavioural health behavioural health data by ensuring appropriate levels of in-house
care services and address health expertise, IT-infrastructure, compliance, and legal requirements to
inequities in rural America. enable new data driven behavioural insights.

16
Digital Healthcare
Adopt
Healthy Human

Digital tools enable completely new

approaches to health and wellbeing.
Today, visiting a doctor remains a burdensome challenge across the globe, especially in
highly developed countries. The journey from the onset of the disease, to visiting a
general practitioner, and then receiving the final diagnosis from a specialist can take
several weeks – which shows some obvious potential for improvement.
Opportunities Risks
Digital healthcare tools can speed up such processes massively. First and initial checks  Patients receive their intervention  The use of personal health data
can be provided on-demand via cell phone (e.g., symptom checker or skin cancer earlier, medication and treatment generated by digital healthcare tools
screening), booking tools support scheduling necessary appointments or send reminders errors can be reduced, and the is limited by different regulations
for regular check-ups, vaccinations or screenings. standard of care increases through worldwide
All of these tools generate valuable and important data which can be stored in Electronic greater continuity  The creation of real value for patients
Health Records (EHR). EHRs allow medical professionals direct access to a patient’s  Incorporating digital healthcare requires the integration of various
complete medical history, test results, treatment record, allergies and many more data. technology and tools across the data sources which – again – requires
With that, EHRs do not only enable faster and better diagnosis but, in conjunction with various steps of the patient journey structured data and high computer
other digital tools and technologies, it also allows new opportunities in the space of generates vast amounts of connected power
disease surveillance and biometric monitoring. data, which enable new and in-depth  Personal healthcare data is highly
analysis and data-driven decision sensitive and must be well safeguarded
Altogether, Digital Healthcare tools simplify the coordination between the different
making
service providers: doctors, hospitals and insurance companies. The data (once  The clinical acceptance of efficiency and
standardised and structured) from solutions such as Diagnosis@Home or  The use of digital healthcare tools sufficiency of digital tools needs to be
Treatment@Home provide valuable data-driven insights in the future, increases process efficiencies while proved over a longer duration
thereby improving existing data models. reducing administrative burdens
 All players in the healthcare system
which can result in potential large
need to be educated and trained to
cost savings and improved patient
use digital tools, especially medical
satisfaction
professionals

17
Digital Healthcare
Adopt
Healthy Human

Dragon Medical One Dynamix Projected Digital Health Market Size CAGR: 18.6%
Source Source
Nuance Communications, Inc. Inspirata, Inc. US$ 211 billion
in 2022 US$ 809 billion
Conversational AI improves Oncology specific AI furthers
clinical productivity pathological diagnostics
Nuance is one of Microsoft latest Inspirata’s Dynamyx software is about
acquisitions, costing US$ 20bn and streamlining digital pathology case The digitisation of the healthcare systems have made significant progress and
marking a massive strategic move into management. projections show that the digital health market may nearly quadruple to a financial
healthcare for the company. It features a fully integrated workflow volume of US$ 809bn by 2030. Also, large government initiatives such as in India or
Dragon Medical One uses advanced tool, an evidence tray that collects all the France are driving this development further.
speech recognition coupled with voice- digital interactions pathologists have
enabled workflow capabilities to create with the case, an image utilisation tool
notes for clinicians and nurses through and an open interface to any scanner or
transcription, also into EHRs. laboratory information management
system.
Key Insight
The tools capabilities now extend to
further clinical tasks such as order entry, Dynamyx was approved by the U.S.
chart search and review, and note retrieval. Food and Drug Administration in 2022,
In a case study with Oxford University allowing customers to use whole-slide
Large investments underline how Digital Healthcare
Hospitals, clinic letter turnaround images for primary diagnosis rather is rapidly integrating into the medical field and
times could be reduced from 12 to 3 days. than traditional glass slides. Fujifilm transforming it.
Clinical organisations hope that these announced in December 2022 that it
improvements will enhance clinician will acquire Florida-based Inspirata Inc. The options for transforming and speeding up processes
productivity, reduce burnout rates as well in 2023. is massive and will impact every stakeholder within the
as costs, and improve overall efficiency next decade.
while freeing up time and resources for
patient facing tasks.
18
AI Medicine
Trial
Healthy Human

Artificial Intelligence supports medical Opportunities

 Improving diagnostic accuracy and
professionals in diagnosis and treatments reducing the risk of missed or
incorrect diagnoses
AI Medicine is a field that uses machine learning and artificial intelligence techniques to  Patient data can be analysed to
analyse medical data and make recommendations or predictions about patient care or recommend the most appropriate
drug development. These systems can be used to support a variety of medical tasks, treatment options based on the
including diagnosis, prognosis, treatment planning, and monitoring of patient health. patient's individual characteristics
For example, predictive analytics systems can be employed to identify patients who are and circumstances
at high risk of developing a particular condition or disease, based on their medical  The potential to significantly improve
history, demographics, and other factors. Complementary to this, diagnostic systems the speed, efficiency, and effectiveness
that use AI to analyse medical images or other diagnostic data to identify abnormalities of the drug development process can
or disease have been maturing regarding early cancer detection employing, both bio- support the development of more
marker-based and visual-based-data. targeted and effective treatments for a
Also, AI-based drug discovery incorporating large amounts of data, such as genomic wide range of diseases
data, chemical compounds, and clinical trial results, can be used in order to identify  By automating certain tasks, such as
Risks
potential drug candidates and predict their effectiveness. data analysis and decision-making, AI  In case the data used to train the AI
This can speed up the drug discovery and development process that currently takes systems can help to reduce the system is biased, it may produce biased
about 10 years by automating tasks, such as data analysis and molecule design, and by workload of healthcare providers and or unfair results, also the system may be
identifying patterns and relationships that may not be immediately apparent to humans. free up their time to focus on other less accurate at diagnosing individuals
AI can also help to reduce the cost of drug development that averages out to about important tasks from other patient groups
US$ 2.6bn for bringing a new drug to the market by identifying promising drug  Vulnerability to cyber attacks could
candidates more quickly and efficiently. compromise patient data and disrupt
healthcare services
 There are ethical concerns surrounding
the use of AI in medicine, including
issues related to patient privacy
and autonomy
19
AI Medicine
Trial
Healthy Human

AI augmented multiomics Partnership for AI-based

analysis platform breast cancer screening
Source Source
Freenome Holdings, Inc. Google Health, Google LLC

Early cancer detection Mammogram Screening

Multiomics refers to the study of multiple In November 2022 Google announced its
"omics" (such as genomics, proteomics, first commercial agreement for their AI
metabolomics) in order to gain a more technology to iCAD, a medical technology
comprehensive understanding of company that provides breast cancer
biological systems. detection services to health care facilities
Freenome developed early cancer around the world.
detection blood tests using a In an earlier study published in Nature,
comprehensive multiomics analysis Google’s algorithm for mammograms
platform that combines expertise in Key Insight performed better than radiologists in
biological and computational sciences to logging fewer false positives and false
look beyond tumor signals with a machine negatives in reading the mammograms
learning classifier in order to detect the Growing regulatory openness in AI drug discovery from more than 91,000 women in the
body’s early warning signs of cancer. European regulation ICH M7 EU facilitates quantitative structure-activity US and the UK.
In 2022 Freenome announced working relationship (QSAR) assessment of toxicity. This computational method is able While the algorithm is not designed to
with the Oracle Cerner network, launching to provide a faster and more accurate analysis of the relationship between the replace radiologists in the near term it
a multiomics oncology study. chemical structure of a compound and its biological activity. aims at complementing the radiologist,
This precision medicine initiative will use since for now, no regulatory agencies
At the same time, the U.S. Food and Drug Administration (FDA) recognizes have signed off on an entirely AI-based
real-world data from the Learning Health single-group trials in rare diseases with control groups incorporating real-world
NetworkTM for clinical validation and interpretation of mammograms.
evidence (RWE).
refinement of cancer risk prediction models.
20
Biomanufacturing
Trial
Healthy Human

Traditional biomanufacturing which has been common  Reduction of the environmental impact
of biomanufacturing processes, for
since the late 1970’s comprises the use of biological example by minimising waste production
processes to produce pharmaceuticals, industrial enzymes, or using renewable resources
 Growing bioeconomies and their use
food additives, chemicals and other products. It involves of novel technologies, built upon bio-
the cultivation of microorganisms or cells to produce a manufactured components, need
specialized business risk solutions
specific product, which can be a chemical compound,
protein, or other biomolecule. Risks
So-called advanced biomanufacturing refers to the use of cutting-edge technologies and  The public may have concerns about
techniques, such as advanced analytical tools and automation, as well as the application the safety and ethical implications of
of engineering principles, to optimise the production, biological processing and design of biomanufacturing, making it difficult
biomanufacturing processes. for companies to secure public support
for their products and processes
Advancements in biomanufacturing have meanwhile led to fuels, chemicals, and
engineered meats and plants, for example, being made from renewable resources.  Regulatory requirements are complex
and time-consuming to navigate.
Subsequently, foodtech companies have lately been focusing on lab-grown meat, which
Ensuring compliance with these
may drastically lower the future carbon footprint of meat production:
requirements can be a significant
Using a biopsy from an animal, scientists isolate high-quality cells from different parts of challenge for companies
the tissue. These cells are then fed with nutrients so they can grow efficiently under
 The development and implementation of
controlled conditions. Finally, they are put in bioreactors, where they grow until ready to
biomanufacturing technologies can be
be harvested and made into meat components.
Opportunities expensive and may involve significant
Concurrently, overarching biomanufacturing-based efforts are underway to develop upfront investments, which could make
“circular bioeconomies”, where agricultural resources are harvested, consumed, and then  Optimising the performance of
it difficult for companies to recoup
regenerated in a sustainable manner into other usable materials, such as sustainable biomanufacturing processes, which
their investments
asphalt and bio-plastics. can lead to higher productivity and
lower costs

21
Biomanufacturing
Trial
Healthy Human

FDA clears lab- US President signs which contains three core

recommendations: to build up
grown meat off executive order biomanufacturing infrastructure hubs, to
to advance streamline the necessary regulatory
Source
approval processes, and to prepare a
FDA biomanufacturing new, data-based strategy for the
bioeconomy.
Lab-grown meat considered safe for Source
human consumption The Whitehouse
After Singapore was the first nation to
allow the sale of cultivated meat to The Executive Order 14081, “Advancing
consumers in late 2020, the US were the Biotechnology and Biomanufacturing
next to follow as the FDA approved food Innovation for a Sustainable, Safe, and
made from cultured chicken cells as safe Secure American Bioeconomy” was
human food in late 2022. signed off by President Biden. Some of
the goals of the executive order are to
Before this food can enter the market, the
facility in which it is made also needs to
bolster and coordinate federal Forecast for the
meet applicable regulatory requirements.
investment in key research and
development, improve and expand
worldwide cultured Key Insight
This includes manufacturing and
domestic biomanufacturing production meat market size
inspection requirements by FDA, FSIS
and USDA.
capacity and processes, and promote 2026–2030 Regulatory and executive
standards, establish metrics, and
Although the FDA has only approved CAGR: 34.8% frameworks take shape
develop systems to grow assess the
products from the food technology state of the bioeconomy. This executive The decisions by the FDA and the
company Upside Foods for use, it has order also set in motion the so-called US government lay the foundation
stated that it is prepared to collaborate PCAST council to release a report in US$ 28.6bn for upcoming biomanufacturing
with more businesses creating cultured December 2022 labelled “Strengthening investments and developments
in 2026 US$ 94.5bn
animal cell food and production methods, Biomanufacturing to Advance the that will require new covers by
including those generated from cells Bioeconomy” their insurance partners.
from seafood.
22
Personalized Medicine
Trial
Healthy Human

More effective treatments and therapies

based on patients’ individual
characteristics
The goal of personalized medicine is to provide the most effective treatment for each
patient, rather than using a one-size-fits-all approach. Opportunities Risks
There are several ways that personalized medicine can be used in clinical practice. One  Potential to improve patient outcomes  Personalized medicine can be expensive
example is pharmacogenomics, which involves using a patient's genetic information to by providing more targeted and and comprises specialised treatments
predict how they will respond to certain medications. This can help doctors choose the effective treatments. This approach that may not be covered by insurance.
most effective and safest medication for each patient. can also help to reduce the risk of side This could create barriers to access for
Another example is the use of genetic testing to help diagnose and treat medical effects and improve the overall quality some patient groups
conditions. Genetic testing can be used to identify genetic mutations that increase a of care and in the end reduce costs  The collection and analysis of sensitive
person's risk of developing certain diseases, such as cancer. This information can help  Development of new treatments that personal data, such as genetic
doctors tailor preventive measures, diagnostics and in case of disease also treatment are tailored to the specific needs of an information, which raises concerns
plans to each patient's specific needs. individual. This can help to address about privacy and the potential for
A lot of progress has recently been made in gene therapy recently. One form of this unmet medical needs and improve discrimination
therapy is called “base editing” which allows scientists to “zoom” to a precise part of the patient care  There is currently a lack of
genetic code and then alter the molecular structure of just one base, converting it into  Help to identify individuals at high risk standardisation and regulation in the
another and changing the genetic instructions. This allows researchers to make highly for certain diseases or conditions and field of personalized medicine, which
specific changes to the genome that can potentially correct genetic mutations or alter provide personalised prevention could lead to inconsistencies in the
the function of a gene, thereby redirecting the immune system to fight cancer, strategies to help reduce the risk of quality and accuracy of tests and
for example. developing those conditions treatments
In this segment there are already trials of base editing under way in blood disorders like
sickle-cell disease and beta-thalassemia, as well as high cholesterol that runs in families.

23
Personalized Medicine
Trial
Healthy Human

Gene-based First use of a revolutionary new type of medicine

The current revision of our cancer underwriting
treatment for Just six years after the technology of base editing
was invented, it saw its first application in 2022 guidelines reflects the continuously improving mortality
Leukemia
where a large team of doctors and scientists used it and morbidity of cancer patients. Personalized medicine
Source to engineer a new type of T-cell that is capable of will help to continue this success story, especially for
NHS hunting down and eliminating the patient’s
cancer types or diseases that are currently difficult to treat.
cancerous cells by redirecting the immune system
to fight the cancer.
This experimental therapy – which contained millions Alban Senn, M.D.
of the modified cells – subsequently resulted in Chief Medical Officer Munich Re
freedom from cancer up to now.

FDA approves New gene therapy approved Key Insight

most expensive “Hemgenix” is a gene therapy to treat adults with
treatment hemophilia B, a genetic bleeding disorder in which
people do not produce a protein needed to create Strong projected growth of global precision
ever for blood clots. medicine markets.
Hemophilia B Manufacturer CSL Behring set the price at The global non-oncology precision medicine market is forecast
US$ 3.5 million per treatment, making it the most
Source to more than triple from 2019 to ~US$ 130bn by 2030.
expensive drug in the world. Despite the high cost,
FDA
the U.S. Food and Drug Administration (FDA)
& CSL Behring LLC
approved Hemgenix based on safety and
effectiveness evaluated in two studies.

24
Population Health
Trial
Healthy Human
 Improve the overall health of a
Population health focuses By considering these factors, which –
community while concurrently reducing
today – are often not collected neither
on the health of a analysed, the prevention of diseases
healthcare costs by addressing the root
causes of poor health outcomes
population as a whole could be more targeted to specific
 Based on population health data, health
people of risk and therefore decrease
to indicate the health current health inequities. insurance services may be modified to
fit the requirements of particular
risk of individuals. Collecting, summarising and analysing subgroups and individuals
such a massive amount of data is only
It involves understanding and addressing
possible if all stakeholders in a health
the various factors that can impact health,
system collaborate - public health Risks
such as social, economic, and
agencies, healthcare provider,  Accurate and timely data is critical for
environmental factors.
community organisations and patients the success of population health
themselves. Lastly, regulators are key as initiatives. However, collecting and
clear and transparent rules and analysing data from multiple sources
guidelines are required. can be challenging, and may require
specialised skills and resources
Population health strategies will lead to
drastic changes in health systems but,  The way care is delivered must evolve
in the end, it supports delivering care in for population health. It involves
a proactive manner as opposed to the coordinated work based on
conventional reactive approach. interprofessional collaboration, and its
complete implementation might take
long time
Opportunities  It needs to be ensured that population
 Help to reduce health disparities health data is collected, stored, and used
among different groups of people, in a secure way that is compliant with
such as those with lower income or relevant laws and regulations and
those living in underserved areas protects the privacy of individuals

25
Population Health
Trial
Healthy Human

Genomic dataset of nearly The Population Health

100,000 whole genome Information Tool
sequences published Source
Source Massachusetts Department of
National Institutes of Health (NIH) Public Health (MDoPH)

The All of Us Research Program of the The vision of the Massachusetts

US National Institutes of Health currently Department of Public Health’s
provides access to about 100,000 extremely Population Health Information Tool
varied whole genome sequences. (PHIT) is to prevent illness, injury and
premature death, to ensure access to
This diverse set of participants can
high quality public health and healthcare
change the landscape of genomics
services as well as to promote wellness
research
and health equity for all citizens.
A little over 50% of the data comes from
To achieve that the MDoPH is building a
people who identify with ethnic groups that
cloud-based tool to aggregate available Key Insight
have historically had lower research
representation. health data across the whole country.
The tool combines data on so-called
By addressing previously unanswered Population health has the potential to evolve health care
social determinates of health such as
concerns about health and illness, this
conditions in which people are born, systems from a therapy-based approach to a preventive
data will allow researchers to advance
grow up, work, live, and age with approach
knowledge and make new data-driven
community health data.
discoveries that will help to close long- This is best supported by a data-driven approach with a structured
standing health inequities. These community health datasets
analytics process driven by skilled and competent healthcare
already include more 14 segments such
The genomic data is available via a cloud- technicians, clinicians and data-analysts.
as infectious diseases, hospitalizations
based platform, the All of Us Researcher
and a cancer registry.
Workbench, which additionally includes
genotyping arrays from 165,000 participants.
26
Robotic Health
Trial
Healthy Human

Robotic health refers to the improve especially minimally invasive  In face of an ageing population and a
surgery, such as cardiac, digestive, lack of work force, elderly care is one of
use of robots and robotic the markets most in need of valuable
gynaecology, plastic reconstructive
technology in the healthcare surgery, throat surgery, neurosurgery, digital solutions and support. This may
industry, such as in surgeries, and vascular surgery. create new opportunities for insurers to
offer new products for robotic care
rehabilitation, and assistance In these examples a surgeon controls  Robotic health systems can monitor
with daily living activities for the robot to make smaller, more precise patients providing real-time information
patients with mobility movements than are typically possible on their condition and alerting health-
with a human hand. Delicate, flexible care professionals if there are any
impairments. changes
movements during complex operations
This can include surgical robots, are made, thereby helping ensure
exoskeletons, and telepresence robots robotic surgeries are easier and Risks
that allow doctors to remotely consult more successful.
 In case robots begin making medical
with patients. The goal of using robotics
decisions or offering unsupervised
in healthcare is to improve patient
outcomes and quality of care, while also Opportunities assistance there is a potential risk
of liability
increasing efficiency and reducing costs.  Robotic-assisted microsurgery allows
 Human error in programming or
Next to hydraulic systems, machine for greater precision, control and
operating the robot as well as mis-
learning, cognitive automation, natural- reduces the risk of human failure as
communication or misinterpretation
language processing and conversational well as necessary corrective
of data may lead to unintended
interfaces are key technologies in this procedures
negative outcomes
enabling robots to interact more on an  In remote surgery, medical
 Robotic health is interdependent on
emotional level by being, for example, a professionals may carry out
other technologies such as computer
robotic companion to the elderly, procedures regardless of where they
vision, sensors or machine learning, all
are. Patients located in remote areas
making them feel less lonely while of which are prone to their respective
could obtain expert-led surgery if they
monitoring their health (e.g., blood own domain-specific errors
need specialized care
pressure, blood sugar levels).

In addition, robots are continuing to

27
Robotic Health
Trial
Healthy Human

Mako Smart Robotics™ for One Versatile humanoid service robot

Total Knee & Total Hip autonomous Aeolus Robotics introduced “aeo”, a dual-arm humanoid robot that can perform several service tasks,
including delivery, security, eldercare, and ultraviolet germicidal disinfection. The robot’s dual arms have
replacement robotic seven degrees-of-freedom, which allow it to perform simultaneous tasks such as delivery or disinfection
platform, two with one arm, while the other arm is free for mobility tasks such as operating elevators or opening doors.
Source
Stryker Corporation
arms, four The robot can also be operated remotely, when necessary, from a smartphone.
areas of Aeo’s care solution aims at maintaining a safe environment for patients and residents while reducing
Robotic-Arm assisted surgery for application staffing needs. Its vision and artificial intelligence enables it to detect, for example, whether hospital
orthopedic implants patients, or the elderly, have fallen or need medical attention. The robot is intended to complete simple
Source tasks that free up staff to have more time to interact with patients such as performing night shift patrol for
Through CT-based 3D modeling of the
unique bone anatomy, surgeons can use
Aeolus resident safety, delivery of medical supplies across hospitals or disinfecting facilities. Aeo is deployed in
Robotics, Inc. Japan, Hong Kong, and Taipei and made its North American debut at the CES® exhibition in
the Mako SmartRobotics™ system to
Las Vegas in early 2023.
create personalized surgical plans that
consider disease severity, joint alignment
and the surrounding bone and tissue. Its
scope of application is for surgeries of the
hip and knees, which includes both partial Key Insight
and total knee replacements and hip
replacements.
Market adoption of robotic health is highly dependent on local
The precise data created at the time of
healthcare systems and their ability to invest
surgery allows the surgeons to continually
assess the procedure and refine the surgical While robotic health has made headway into different medical spaces,
plan intraoperatively if necessary. The Mako high initial costs, limited availability, and differing legislation usually are
robotic arm allows for greater precision factors limiting its further progression.
during the operation, and that means fewer
patients need follow up surgeries as it is
often necessary to fine-tune the orthopedic
work done in previous surgical procedures.
28
Human Enhancement
Assess
Healthy Human

Human enhancement refers to Their overarching potential human Risks

application aims at restoring or
age-old human desire to improve  The long-term impact of such improve-
regaining parts of body functionality
or augment human physical, or of sensory function, such as vision, ments could be difficult to predict.
cognitive, or psychological hearing or the control of prosthetic Significant health risks and related
expenses for insurers might result
abilities beyond their natural limbs.
levels using technology or  There are complex legal and regulatory
challenges in order to ensure safety
other means.
Opportunities and fairness of such enhancements
Rather than simply treating disease or  Potentially improving physical abilities  People may feel pressure to enhance
injury, the goal of human enhancement is such as strength, endurance, reflexes themselves in order to keep up with
to enhance or augment human capabilities. and mobility. This could have a others or to meet societal
Examples include using prosthetics to significant impact on people with expectations
improve physical abilities or using brain- physical or cognitive disabilities,
machine interfaces to enhance or regain allowing them to perform activities
abilities. The application can range from that would otherwise be impossible
permanent physical modifications of the
 Augmenting research on the brain and
human body to regain healthier living, to
neural functioning, which could lead
wearable gadgets with the key goal of
to new insights into the workings of
enhancing a healthy human’s capabilities.
the brain and new treatments for
Recently so-called brain-machine interfaces brain disorders
(BMIs), which are systems that allow  Regaining physical abilities: prosthetic
communication between the brain and an could be used to replace lost limbs;
external device, along with brain-computer cochlear implants can be used to
interfaces (BCIs), which allow a person to restore hearing and/or exoskeletons
control a computer or other device with could enable people to walk again
their thoughts, have made progress towards faster after a stroke
the commencement of initial clinical trials.
Human Enhancement
Assess
Healthy Human

The Science Eye In a three-step procedure the implant is Bypassing damaged superior sagittal sinus using a neuro-
first slid in over the eye and the display interventional procedure. The BCI
Source is inserted through a small incision. neural pathways with a appeared to be safe and enabled patients
Once the implant is secured in place to do hands-free texting, emailing,
Science Corporation BCI implant online banking and shopping, and
and the display is laid on the retina, the
communicating care needs using their
“Prosthetic eye” combines implants, display film is affixed to the retina to Source
thoughts.
gene-therapy and wearable hardware hold it in place. Synchron Inc.
The company is working on The study indicated that it is possible
In late 2022 the “Science Eye”, an
First in-human clinical trial completed to record neuronal signals from a blood
advanced optogenetic visual prosthesis demonstrating early functional results
vessel in the brain as well as a favourable
device, was announced by the Science of the device with animal trials, trying to The Syncrhon Stentrode™ is an safety profile of this approach.
Corporation. This device could offer new produce sufficient safety data, so that a endovascular implant, placed in a vein
hope for patients with serious blindness first-in-human clinical study could alongside the motor cortex which
due to photoreceptor loss attributable to potentially be started. generates any signal related to
retinitis pigmentosa or macular movement. A receiver device implanted
degeneration, for example, by restoring in the chest transmits the neural signals
visual input to the brain. to a decoder. Finally, a machine learning
algorithm translates those signals into
The Science Eye employs a combined
specific digital commands. Key Insight
approach. At first optogenetic gene
therapy renders the optic nerve cells light- “The Stentrode With Thought-
sensitive at a specific wavelength by Controlled Digital Switch (SWITCH)
adding a protein to the cells of the optic Study” published early 2023 questioned
Advancements in
nerve. Then a surgically inserted high- if an endovascular brain-computer neuroprosthetics
resolution microLED display film controls interface (BCI) implant is safe, and if it
Wider and more rapid
these newly light-sensitive ganglion cells can enable paralysed patients to control
together with the corresponding implant. translation of BCIs to people
a computer with their thoughts. In this
Finally, glasses worn by the user contain case series including four patients with seems possible, expanding
sensors, processing, and batteries, severe bilateral upper-limb paralysis, the insurance requirements.
supplying power and data to the implant. BCI device could be implanted into the

30
Vaccine Technologies
Assess
Healthy Human

Vaccines will enable new treatments

of severe and rare diseases
mRNA vaccines are a type of vaccine that use a small piece of genetic information called
messenger RNA (mRNA) to stimulate an immune response. Their first broad application
was fueled by the COVID-19 Pandemic during which multiple mRNA vaccines were
developed, clinically trialed, approved by the corresponding legislation, and subsequently
distributed world-wide with a never-before-seen speed.

In contrast to traditional vector vaccines, mRNA-based vaccines work by providing the

body with the instructions to produce a specific protein or fragment of a protein that is
found on the surface of a virus or bacteria.

The advantage of mRNA vaccines is that they can be developed and produced quickly,
as they do not require the use of live or attenuated viruses or bacteria. In addition, mRNA
vaccines have a good safety profile, as they do not alter the genetic material of the host Opportunities Risks
cells and are rapidly broken down and eliminated from the body.  The design and production can flexibly  The general public’s acceptance of
target a wide range of infectious agents, vaccines is essential. One of the primary
Based on the highly successful development and application of mRNA based COVID-19
including viruses, and potentially reasons why measles has not been
vaccines this technology is being expanded to other areas of application, for instance
bacteria, and parasites completely eradicated yet, for example,
BioNTech announced in late 2022 that it aims to develop the first mRNA-based vaccine
 They are versatile and can be tailored is the criticism surrounding the
for Malaria prevention based on a novel multi-antigen vaccine approach. Today many
to specific proteins or fragments of vaccination based on potential
further initial mRNA-vaccine tests and trials are in the pipelines of pharmaceutical
proteins that are unique to certain side effects
companies, aiming at developing vaccines against specific types of cancer as well as
strains of a virus or bacteria  Access to all people who need vaccines
other diseases and viruses.
 mRNA vaccines can be developed and can be a challenge, particularly in low-
produced more quickly than other types income countries or in areas with limited
of vaccines, as they do not require the use healthcare infrastructure
of live or attenuated viruses or bacteria
31
Vaccine Technologies
Assess
Healthy Human

Investigational Prophylactic Herpes

personalised mRNA Simplex Virus-2 vaccine
Melanoma vaccine trial trial Every step, every patient we treat in our cancer trials, helps us to
find out more about what we are against and how to address
Source Source that… We have a number of breakthroughs, and we will continue
Moderna, Inc. & Merck & Co Inc. BioNTech SE to work towards our goal. To translate science into survival.
Primary efficacy endpoints met First In-human study initiated
In December 2022 MODERNA and In December 2022 BioNTech announced Prof. Özlem Türeci, M.D.
MERCK announced the first demonstration that “BNT163”, the first candidate from
of efficacy for their investigational mRNA- their infectious disease mRNA vaccine Founder and Chief Medical Officer at BioNTech
4157/V940 cancer treatment in a collaboration with the University of
randomised clinical trial. Pennsylvania, will enter the clinical trial
The results of this randomised phase 2b in order to evaluate the safety and
trial demonstrated a statistically significant immunogenicity of this Herpes Simplex
Virus vaccine candidate.
and clinically meaningful reduction in
Key Insight
the risk of disease recurrence or death
compared to the corresponding
monotherapy. Initial mRNA vaccine trials
This was the first time ever, the potential are rapidly expanding
for mRNA to have an impact on outcomes
From Malaria and Herpes to
in a randomised clinical trial in melanoma
Melanoma and multivalent
was demonstrated. Both companies plan
Influenza, initial mRNA vaccine
to discuss results with regulatory
tests and trials have gained a lot
authorities and to initiate a phase 3 study
of traction during 2022.
in melanoma 2023 thereafter planning to
rapidly expand to additional tumor types.
Our Solutions
Healthy Human
Munich Re Munich Re

Claudia Allen MIRA PoS MIRApply Insured

Product Owner Brings our know-how to your Improve client interaction with
MIRA PoS and point of sale digitalised queries
MIRApply Insured MIRA PoS is a “software as a service” Once the insurance application details
product developed by Munich Re which provided by a potential client have been
digitalises your processes at the point checked, further information (such as
of sale – from the capture of case and relating to pre-existing conditions or
application data, via risk assessment pastime risks) may be required. This
Integrate Munich Re’s to reporting. necessitates the underwriter going back
What’s so special about this MIRA to the client to gather additional details.
underwriting expertise Digital Suite module? It works invisibly MIRApply Insured transforms the
directly into your in the background for applicants, gives traditional laborious process, making it
you plenty of room for differentiation faster, more accurate and more client-
applications. from competitors and, depending on friendly. The web-based tool virtually
the scope of use, delivers risk-related or eliminates the need for time- and paper-
Digitize the application even case-closing acceptance decisions intensive requests, making the whole
that are based on the very latest process markedly more efficient.
enquiry process. knowledge.
Our Solutions
Healthy Human
Munich Re
CLARA plus
Stephan Jacob For the first time, a system-
Product Owner based calculation of the
CLARA plus degree of disability
Until recently, it was common under-
standing that there is no way to
digitalise the claims handling process
for disability insurance. The traditional
Support the claims process, which calculates the degree of
disability manually, is time-consuming
decision process and requires in-depth medical expertise
in a structured, and is therefore less efficient. With the
introduction of the CLARA plus service,
consistent way. Munich Re now empowers its clients
to calculate the degree of disability
automatically, leveraging the
following benefits
 Enhanced resource management
 Cost reduction for claims
administration and processing
 Transparency and reproducible
decision-making
 Improved reactivation rate
 Data analysis
Our Solutions
Healthy Human
Munich Re Munich Re
Predictor ALLFINANZ Spark
Artificial Intelligence – Life insurance’s new generation of digital underwriting
Use of predictive models in Munich Re Automation Solutions help life insurers all over the world to create up-to-
Life underwriting date solutions for different parts of the insurance value chain.
Life insurers looking to evolve beyond For example, ALLFINANZ SPARK is a new cloud-based business, digital life insurance
underwriting and analytics solution designed to deliver everything an insurer needs
simple underwriting rules and combine
quickly. With its automated underwriting system and capabilities, life insurers can
their traditional tree-based approaches
reduce many of the routine cases that take up so much of their underwriters’ valuable
and emerging AI predictive models, can time. With access to new sources of data and insights, you can make better decisions
accelerate their underwriting innovation much faster. Improving the efficiency of automated underwriting processes through
with Predictor from Munich Re digitalization and the use of artificial intelligence, the next step – moving to augmented
Automation Solutions. automated underwriting – is not far away.
Predictor allows forward-thinking
insurers to operationalise, monitor and
manage predictive models within their Munich Re
underwriting process. It removes many Insight
of the barriers that have prevented
Data integration and decision-making in the Life underwriting process
carriers from adopting predictive
underwriting to date. It opens a clear Life insurers looking to exploit their data resources and to operationalise data-driven
path to lower referral rates, reduced decisions within their business use Insight Analytics to make their data accessible,
evidential burden and improved with targeted analytics aligned to business goals.
customer journeys. Insight Analytics provides a deep and wide view of underwriting processes, data and
technology, combining raw data from multiple sources, enriching it with structure,
and making it accessible to the business. It supports customers at all stages of their
analytics journeys, from “out of the box” capabilities to curated data that can be
leveraged by business users to build, share, interrogate and collaborate on data
insights relating to their business.

35
Our Solutions
Healthy Human
Munich Re Munich Re
IQ2Health SMAART and Visualizer
Designed to streamline the A comprehensive toolset to
underwriting process while monitor, steer and grow the
enhancing speed and consistency health portfolio
A continuous work in progress, IQ2health Specialized portfolio steering tools
is designed to support an A-to-Z solution. developed by Munich Re HealthTech
From receiving a health quote to issuing to provide key insights which can be
a quotation, IQ2Health goes through the leveraged by our clients. The toolset
automated process of individual and includes
group quotations for new and renewal  Portfolio Tool – Monitor KPIs of
business (for all product types). entire book of business
Benefits at a glance  Group Tool – Analyse and renew
 Automated medical underwriting group policies with all the details
decision-making based on a member’s  Individual Tool – Analyse and renew
medical history for new and individual polices
renewing policies  Policy Performance Monitor –
 Automated administrative/ Detailed analysis at policy level
technical underwriting decisions for  Visualizer – a segmented portfolio
new and renewing policies approach to enable a deep
(e.g., enrolment eligibility) understanding of performance to
 Automated risk pricing devise and adjust an optimal profit-
(premium setting) making strategy
 A consistent and faster  Key alerts and notifications
underwriting process
 Higher hit ratios and lower
operational expenses
Connected Experience
Technology impacts in 2023

Around The edge computing

4.4bn market is assumed to grow
Until 2023 at a CAGR of
connections with 5G
is estimated
by 2027.
2.4m
subscribers will use
Low Earth Orbit satellites
38.9%
services by
The Asian 2023. The global
Pacific market is the
fastest growing in 60% data fabric market
Is estimated to grow to
brain computer of published automotive

US$ 6.97bn
IoT were conducted
interface by friendly
market. researches.
within the next
7 years.

38
All trends 2023
Connected Experience
The trend field of connected experience comprises trends related to or enabled by network and One of the interesting trends in 2023 is ambient computing – the ability to use hardware and
infrastructure technologies. We speak of a connected experience when various systems, devices and technology unconsciously. Ubiquitous connectivity describes the continuous connection with the
entities interact with each other e.g., through data fabric or IoT infrastructure. With the technology to world independent from location and time. As of today, there is ongoing development around 5G and
help reduce data or be able to process it more energy- and cost-efficiently, connected experiences its successor technologies, but in the future, self-driving cars or facial recognition systems could
become a prerequisite for building digital infrastructures. benefit from these standards.

Adopt – Start initiatives in your unit Trial – Initiatives in affected units Assess – Evaluation needed Hold – Watch list
Automotive IoT Ambient Computing Brain-Machine Interface
A complex system of devices like sensors, cameras or GPS Combining Smart Home and Internet of Things to use A brain-machine interface interprets the user’s
tracker make long- and short-range wireless connections computers and the Internet without consciously using brain patterns to control external software and
which generate enormous amounts of data. them. hardware.

Data Fabric Extended Reality Metaverse

Fabric-like data management infrastructure Real, augmented and virtual worlds are increasingly As part of the developments of the “Internet in 3D”
for distributed networks that create unified data merging and changing people’s perception of Metaverse advocates a multitude of interoperable
environments to share formerly disparate, their environment. platforms on which users can present as avatars, cooperate
inconsistent data. for business or leisure, own assets, and innovate.

Edge Computing Super Apps

Edge computing processes data as close to its source These are a single point of access for various everyday
as possible – at the edge of the network structure. digital customer services, as already successfully used in
Therefore, the application services reduce the volumes some countries.
of data that must be moved.

Ubiquitous Connectivity
Greater amounts of data volume are being processed on
wireless networks and, combined with IoT technology,
these amounts are being processed more energy- and
cost-efficient.

new in 2023

39
Trend Evolution
Connected Experience

Evolution over time

Adopt Trial Assess Hold

Start initiatives in your unit Initiatives in affected units Evaluation needed Watch list

Ambient Computing 2023

Brain-Machine Interface 2023 2022

Automotive IoT 2023

Data Fabric 2023 2022 2021 2020

Edge Computing 2023 2022

Extended Reality 2023 2022 2021

Ubiquitous Connectivity 2023

Metaverse 2023 2022

Super Apps 2023

40
Impacts on the Value Chain
Connected Experience

Explanation Insurance Value Chain

The illustration represents the impact the trend
fields have upon on the insurance value chain. Product Sales & Risk Customer Asset
With the dark boxes reflecting areas of high Design & Pricing Distribution Underwriting Policy Admin Management Engagement Claims Management
impact, and light boxes low impact.

Ambient Computing

Brain-Machine Interface

Automotive IoT

Data Fabric

Edge Computing

Extended Reality

Ubiquitous Connectivity

Metaverse

Super Apps

41
Automotive IoT
Adopt
Connected Experience

Commercial vehicle Opportunities

makers, dealerships,  Develop make/model/year security scoring model, similar to crash
protection scores
motor carriers, and
 Partner with FBI’s Automotive Sector Specific Working Group
operators increasingly (and other subject matter experts such as Automotive-ISAC USA,
Europe, Japan)
add sensors which
 Support business continuity and resiliency planning for fleets
generate enormous
 Provide affirmative add-on motor coverages to dealerships, motor carriers,
amounts of data. and owner/operators for e.g., expenses on towing, diagnostic, repair
and replacement
To share all this valuable data,
automotive sensors make long-
and short-range wireless
connections, including connections
Risks
to devices within or nearby the  Known vehicle exploits are increasing in sophistication and frequency and
vehicle or trailer. are expected to increase in severity. Motivated threat actors can bypass
These numerous connections layers of security such as network segregation and authentication
present attractive avenues for  Remote API hacks require no special hardware and only low-level technical
insertion of malicious code by expertise. In addition to stealing data, exploits can disrupt, lockdown, or
threat actors seeking financial takeover control of individual vehicles, as well as fleets of vehicles
gain or harm.
 20-year vehicle lifespans concern experts who now seek government
Some say we too often trade funded back-stops to pay damages resulting from potentially massive
increased convenience for automotive IoT event which could cost insurers more than $10bn
decreased security.

42
Automotive IoT
Adopt
Connected Experience

Driving the future of Insurance Value Chain

Automotive Security
Research Product
Design & Pricing
Sales &
Distribution Underwriting Policy Admin
Risk
Management
Customer
Engagement Claims
Asset
Management

Industry
Motor
Source
ASRG
Security thanks to community Remotely connecting
The Automotive Security Research Group
(ASRG) is a non-profit initiative that aims
to a “100% trust protocol”
to identify solutions for the security of network is a material
automotive devices, such as software- Physical In-Cab Device Connectivity/Communications
controlled vehicles. The focus lies in
risk to both safety
sharing knowledge and expertise along and security.
with conducting community research
projects to make the roads a safer place.

Tablet/phone App Mobile App

During 2022, only 37% of
publicized automotive IoT Vehicle Connection Cloud or Back-end

exploits were conducted by Network/ Data Store Management

Bill Burrows
Modem Connection Link System
friendly researchers. Munich Re Specialty Insurance
IVG Connected Mobility Risk Solutions lead
43
Data Fabric
Adopt
Connected Experience

A data fabric enables analysis and sharing Opportunities Risks

of formerly disparate, inconsistent data from  Fast provisioning of a data product  High implementation costs for bespoke
distributed networks and different data sources. for insurance use cases, for example
in underwriting
data fabric requires the redesign of static
infrastructure to enable a mesh
 Providing integration pipelines for network environment
fast-growing data generated also by  Strong data security as well as access
partner networks and ecosystems policies and technologies are needed
with a variety of applications to ensure proper sharing of data
 Overcome data silos with limited use  Profound understanding of data
cases to enhanced usage of data management and data governance
based on metadata management is a prerequisite for effective data
Data fabrics are designed to enable applications and tools to access and transform fabric implementation and usage
different types of data, regardless of their storage location in the cloud or on premise,
and supporting established data standards and systems.
They create unified data environments in distributed computing networks – even
among reinsurer, insurer, agent and third-party networks. Due to a metadata driven
approach they provide a solution to the ever-growing data quality problem which is a
key challenge for the successful implementation of all data-enhanced technologies.
Next to core capabilities such as data integration or support of data governance, the
use of knowledge graph technology in a data fabric design, for example, also promises
enhanced business value for different users through the combination and
contextualizing of data.
There are various products and platforms available in the market promising
data fabric solutions, however, independent of the underlying technology, the key
constraint to be tackled is the maturity of knowledge and organizational handling of
data management.

44
Data Fabric
Adopt
Connected Experience

Denodo Platform
CAGR Industry

Universal leverage of data is the order Various

+22.2%
of the day in insurance companies. 2022 2029
Source
Denodo
Typically, it is stalled by huge efforts to Fueled by the ongoing shift into the Logical data fabric
cloud the global data fabric market is
access and integrate data locked in silos estimated to grow within the next seven
Denodo’s solution comes along with a
common semantics layer including
across complex application and cloud years from $1.71bn to $6.97bn. dynamic data catalogue for semantic
search and data governance features.
landscapes.
Pursuing the metadata- and distributed
Key Insight
domain-driven paradigms of the data
fabric is a great approach to unlock data
Data is key in a digitized world.
and make it easily available to numerous Insurers need to move towards a data driven architecture
BI and AI use cases. leveraging the value of customer and insurance
data thus getting advanced business insights.
Georg Opora

Senior Enterprise Architect

ERGO

45
Edge Computing
Adopt
Connected Experience

Edge computing collects automotive, smart home devices, Risks

food and healthcare for example.
and processes data as close  Edge computing needs proper
Edge computing use cases across all
to its source as possible – security to protect local data
industries are expected to grow even
at the edge of the network further over the next few years.
e.g., through transport and
storage encryption
structure.
 The edge algorithm needs to
Transmitting large amounts of raw
distinguish between important
data over a network or into a cloud puts
a tremendous load on network resources. Opportunities and unimportant data. Errors might
lead to wrong business decisions
In many cases, it is much more efficient to  For health care or preventive services
collect and process data near its source like fall detection, connected devices  Supply for local computing power
and send only the data that has based on edge technology can needs to be ensured
overarching business value over support customers
the network.
 Local processing enables a better
Edge application services significantly response to local security
reduce the volumes of data that must be requirements
moved, thereby reducing transmission
 On the edge data processing
costs and improving speed. Due to lower
can proceed even if a (temporary)
latency, near real time services including
loss of connectivity occurs
real time analytics are available.
 In case of claims data on the
The move towards edge computing is
edge, e.g., from IoT devices or robots
driven by mobile computing, the decreasing
are available and could be used for
cost of computer components including
insurance use case handling
storage, and the sheer number of
networked devices.  It could improve the underwriting
of risks based on big data in areas
Along with current network technologies
with slow network speeds
such as 5G, edge computing enables
innovative use cases in industrial IoT,
46
Edge Computing
Adopt
Connected Experience

$11.24 billion
in 2022 $155.90 billion

Edge Impulse Maersk It is assumed that the global edge computing market will
grow at a CAGR of 38,9% within the next 8 years.
Industry Industry
Various Logistics
Source Source
Edge Impulse CIO IDG Communications
ML on the Edge Next generation port Key Insight
Edge Impulse focuses on solutions The Danish company aims at
regarding edge devices with moving data from cloud to the Insurers can benefit from maturing edge computing
embedded machine learning. For edge to support real-time data
instance, developers can import processing and decision making.
solutions in B2B and B2C business, e.g., in telematics,
data from sensors, cameras or By the end of 2023 80% of the smart houses, logistics or supply chain management.
audio devices to embedded cloud data of its port operating Edge computing increases the granularity of data
systems or GPUs. company "APM Terminals" is to points, which should be leveraged for risk assessment,
be migrated to the edge.
customer assistance services and claims handling.

47
Ambient Computing
Trial
Connected Experience

Taking the concepts of smart that an ambient future is in sight. This Risks
is especially true for artificial
home and the Internet of Things  Low maturity or lack of industry
intelligence in the form of natural
(IoT) to a next level, Ambient language processing and large standards inhibits the ability for
Computing promises the use of language models as well as for devices to interoperate
computers and the Internet advanced communication standards.  Poor IoT sensor data could skew
without consciously using them. Ambient Computing's ability to
"get it right“
While the number of smart devices in
homes, enterprises and public settings has
Opportunities  Privacy concerns and consumer
 More accurate risk evaluations based acceptance
increased continuously over the past years,
they often turned out to bring less on additional available data points  Network inconsistency prevents
convenience to users than initially expected.  Predictive maintenance on equipment stable data exchanges
Single smart technologies do facilitate daily to prevent breakdown  Low density throughput from single
operations, yet the target vision of a smooth channel radio do not provide
 Understanding occupancy; e.g., water
and holistic user experience has not enough data
is shut off because nobody is home
been fulfilled.
 IoT fueled Ambient Computing could
Ambient computing aims to close this gap.
reduce the need for visual inspections
In an ambient approach, technology is all
around us and functions deeply  Sensor data could be relayed to
anticipatory: a person’s gaze, movements insurance companies, causing home
and behavior patterns suffice to determine insurance premiums to fall
user needs and to provide a perfectly  Ambient Computing is data driven
personalized service. To reach this level of and will benefit from consistent IoT
usability and direction without human data streams
involvement, the sensors and gadgets need
to be ubiquitous, inter-connected, aligned –
and ‘smarter’ than ever. Recent progress in
enabling technologies give rise to hopes
48
Ambient Computing
Trial
Connected Experience

An ambient approach gets the tech out of your way so you With Ambient Computing, it's easy to envision a future where
can live your life while getting the help you need. your experiences are effortlessly supported by machines and
It doesn’t matter what device you are using, what context
certain risks scenarios are mitigated without human intervention.
you’re in, whether you’re talking, typing or tapping.
The technology in your life works together seamlessly.

Rick Osterloh Norman Thoms

Google’s Senior Vice President of Devices & Services, VP IoT Technologies

Google I/O keynote, May 2022 HSB

Artificial Key Insight

Intelligence

Pervasive
 Meshify, a HSB company, builds IoT If a manufacturing business, for example, can predict
Computing
Ambient technologies that predict and prevent equipment failure based on the information generated by a
intelligence property and equipment risks
network of embedded intelligent sensors and algorithms,
 Its leak and freeze technology is it will be better able to improve its inventory management,
integrated into Amazon’s Sidewalk
Ubiquitous bandwidth sharing network
reduce costs and even prevent potentially harmful failures.
Computing Today, humans are still required to make decisions —
particularly where disparate systems are involved.

49
Extended Reality
Trial
Connected Experience

Real, augmented and With the push in meetings being held

in virtual and mixed set-ups during
Opportunities
 Different loss scenarios could be visually
virtual worlds are the pandemic, now comes the need
represented for the customer, illustrating
for further enhancing communication the value and importance of insurance
increasingly merging, technologies. coverage
The next generation of virtual meetings  In the area of car insurance, extended
changing people’s perception of could be real-time holographic settings reality can provide a range of personal
or communication via special glasses, driving-assistance functions including
their environment. Based on now augmented driving, collision warnings
another move towards merging reality
available 5G technologies and virtuality. Holograms are not new and “black-box” services such as video
extended reality (XR) will become but have lacked broad adoption to date. recording

more common in future. However, research indicates that  In a hybrid mode meetings and trainings
comprehensive holographic can be offered supported by XR
communication might need even 6G technologies, creating one collaborative
environment
network technologies still being in
its infancy. Risks
In the future employees and clients  Acting in virtual or mixed realities can
scattered around the globe will be able to cause physical and psychological stress
sit and interact with one another as reactions. Users need to watch their
though physically co-located in the same mental health and keep a sense of reality
room. Companies will be able to provide  Advanced connectivity and less
new ways to offer training, product dependency on VR-hardware are
information or customer services in more the infrastructural prerequisites for the
natural and relatable ways. adoption of enhanced reality applications

By swiftly addressing customer concerns  Malicious applications could result

in data leakage, such as providing
in that area, the insurance industry has a
unintentional insights into a user’s
potentially significant opportunity for risk
communication, field of view or
assessment and revenue. location, for example
50
Extended Reality
Trial
Connected Experience

Project Feels like you're there

Project Mapping the world

Starline Aria
Project Starline is an attempt to have virtual meetings Project Aria is defined as a research device. It is worn like
with life-size 3D holograms of the participants. This is regular glasses and helps meta employees to conduct
possible due to major research advances across machine research on future Augmented Reality technologies. The
Industry learning, computer vision, spatial audio and light field Industry device will collect data of the viewer's audio, video, eye
display systems. Google employees are already using tracking and local information.
Various Starline and the testing will now go through an early
Various
Currently Project Aria uses LiveMaps. With the help of
Source access program. Project Starline results show that the Source computer vision, it constructs a virtual representation of
Google natural interaction can increase the presence, Meta parts of the world that are relevant to you. The goal is to
productivity and attentiveness of an employee compared create personalized, virtual 3D maps in future AR devices.
to traditional video call meetings.

Augmented Reality (AR)

AR is the technology for the integration Key Insight
of digital information to a user’s real
time environment.
CAGR
Virtual Reality (VR) +36.9% 2021 2028
VR/AR becomes more mature, recognized and used by
customers
VR is a computer-generated 3D
environment that places the user The real and virtual worlds will continue to merge. Along with related
inside a simulated experience. advancing tech trends like metaverse and AI technologies this
It is estimated that the global market
gets pushed even more. For insurers it could improve business
Mixed Reality (MR) for AR/VR/MR will rise from about
processes, communication and collaboration within CRM or
MR describes the merging of a digital $28bn in 2021 to more than $250bn
until 2028. claims management, for example.
world (VR) with a real-environment (AR).

51
Brain-machine Interface
Assess
Connected Experience

A brain-machine Pilots of spelling/writing devices are  Users with limited motor skills
could increase their mobility and
already being conducted by individuals
interface interprets with partial or complete inability to independence by means of such
devices. Manufacturers of medical
the user's brain signals to control their muscles.
devices will be able to apply brain-
control external Current projects aim to extend the machine interface hardware to a wide
scope of these devices to enable greater range of projects. For example,
software and hardware. brain-robot interaction. These techniques bypassing damaged nerves to reconnect
combine brain, gaze and muscle brain and body or replacing conventional
Research and development on brain- prosthetics
tracking to offer hands-free interaction.
machine interfaces (BMI) progress step-  It may also be possible in the future
by-step. Corporations and Universities There are invasive methods where for to combine brain-machine interfaces
also partner in projects for enhancing example electrodes are surgically with VR/AR technologies
and testing promising technologies. implanted in the cortex and non-invasive
approaches to brain-machine interfaces
using scalp electrodes, for example.

However, though there is significant Risk

investment in this field, there are still  In the event of technical failures,
major challenges to overcome before we irreparable brain damage could occur.
see wide adoption, or a significant This applies especially for invasive
market emerge. methods where chips are implanted
into brains, for example
 Using brain-machine interfaces to steer
robots or drones may have unexpected
Opportunities consequences for human users in the
 Brain-machine interfaces to robots long term, such as a loss of the sense
have the advantage that users of one's own body
with limited communication ability  There are fears of privacy invasion and
will be able to control the devices external control
with their mind
52
Brain-machine Interface
Assess
Connected Experience

University of Padova Korean Advanced

Industry Institute of Science
Healthcare Industry
Source Healthcare
Forbes Source
Wheelchair control with your brain Korea biomedical Review
Fully paralyzed patients can control and
Robotic arm controlled by BMI
direct their wheelchair with a cap utilizing
The Korean Advanced Institute of
a non-invasive brain machine interface
Science (KAIST) developed a solution
technology that decodes brain signal
that allows a disabled person to regulate
waves.
a robotic arm prosthesis. This works
thanks to the newly developed Brain-
Brain Computer Interface Market machine Interface that is able to decode
Growth Rate by Region human brain signals. Thanks to cerebral Key Insight
cortical signal decoding the researchers
are able to prognosticate the movement
accuracy of the imaginary arm
extension. The direction of the extension
Further research and advances in the treatment of
can be predicted by up to 80 percent as disabled persons via brain-machine interfaces, for
reported by the researchers. According
High
to the team of KAIST the patient will
example, are pushing this trend. Insurers should
Medium
Low also not need any long-term training to assess the impact that further technological
control the robot arm.
North America is the largest market, development may have towards broader usage.
whereas the Asia Pacific market is
the fasted growing one

53
Metaverse
Assess
Connected Experience

While the ‘metaverse mania’ has calmed down in recent months, Opportunities Risks
developing the ‘Internet in 3D’ remains a key topic on the agendas of  Businesses in general can take  Metaverse initiatives are still in an early
many tech companies and draws attention from users and brands. advantage of the metaverse as an phase with low user numbers, and
The metaverse describes an immersive digital space that blends reality and virtuality interactive cooperation space. The standards are evolving. High
through sophisticated augmented and virtual reality technology. The vision goes beyond metaverse is well-suited for employee commitment to specific platforms or
the gaming worlds that have been known for years: Metaverse advocates foresee a training sessions, collaborative use cases may be premature. Further,
multitude of interoperable platforms on which users can present as avatars, cooperate meetings, or brand representation realizing the metaverse vision will
for business or leisure, own assets, and innovate.
 As ownership of digital assets is require high and ongoing investments
Big technology companies like Meta, Microsoft and Nvidia are at the forefront of the from technology companies. Time will
expected to gain further popularity
movement and have high stakes in related technology, including processing technology, show whether their commitment to the
virtual reality headsets and proprietary metaverse worlds. On the other hand, on metaverse platforms, so will
insurance of these assets. Metaverse metaverse is that persistent
community-driven developments, often rooted in gaming, are building alternative
metaverses that aim to grant governance rights to users. tokens can take on a wide variety of  Bodily harm can result from increasing
purposes – from land to currency to metaverse participation. Accidents may
governance rights. Protecting and occur due to low attention levels in the
insuring these rights, as well as the physical world. Usage may also have
integrity of avatars and metaverse- mental health implications, such as user
related hardware like VR glasses, will addiction or depression
be a priority for many users  As a digital space, the metaverse may
 The simulation capabilities of the be a target for cybercrime and may
metaverse help train for or even attract bad actors that exploit the
replace risky physical activities. For ecosystem. Further, liability risks
example, virtual healthcare and emerge if the metaverse enables
medical training, as well as testing of remote inspection or health treatment
autonomous vehicles on virtual roads  Considerations of metaverse growth
 The metaverse will offer ample and its compatibility with ESG targets
opportunities for behavioural need to be addressed given energy and
data collection hardware requirements

54
Metaverse
Assess
Connected Experience

Technologies that form the Metaverse Financial services via a ANA GranWhale is a metaverse
SAP Metaverse metaverse platform, supported platform operated by the ANA airline
Cryptocurrency by Sompo Japan Insurance company. Users will be able to shop
NFT
Digital Twin
and carry out other activities in the
Distributed
Digital
Mixed Extended virtual world. Sompo Japan is
Ledger Reality
Reality
Identity
Digital Asset (MR)
considering whether they can create
Technology Decentralised
Autonomous
Marketplace Virtual Reality
(MR) a way for users to take our insurance
Organisation
by using MUFG Bank's "Money
Augmented
Reality
Decentralised (MR)
Finance
Mobile AR
Canvas", which is a content that is
located within the Shopping mall in
Web3 Low-Code/ 3D Realtime

the metaverse.
No-Code Rendering
Virtual Platforms
Immersive
computing

The Metaverse Current challenges

Standards Forum To gain adoption and traction, the metaverse needs to overcome various challenges
Clear use case and Privacy and Especially the industrial
 Founded in June 2022, the
value proposition security metaverse, as a form of enhanced
Metaverse Standards Forum
digital twins, has progressed in the
unites leading standards
organizations and companies to past year; corporate and consumer
foster interoperability standards metaverses are not as
for an open metaverse advanced, yet.
 Members include Meta, Microsoft,
Dr. Esther Nagel
Adobe and more than 30 other
technology companies and Business Analyst
organizations Connectivity and Ecosystem
Munich Re
hardware usability interoperability
55
Super Apps
Assess
Connected Experience

Super Apps strive to be the Opportunities

single point of access for  Opening up new customer segments
for insurance services and products
various everyday digital in an embedded Super App
environment
customer services. Already
 Partnering with Super App providers
successful in Asia the might be a cheaper than creating one’s
own app
concept might affect
 Providing easy and low-threshold access
business models worldwide. to digital insurance services in a trusted
environment for sales, service or
In contrast to single-use apps, Super Apps
even claims
offer multiple online services ranging
from banking, medical appointments,  Optimize processes based on API-driven
food shopping to lifestyle. approach necessary for seamless
integration in external platforms
These Super Apps aim to provide a
convenient experience in one holistic
app without the need for the user to Risks
switch between different applications.
Enterprises need to evaluate opportunities  Not exploiting the Super App platforms
and business models thinking about can lead to a loss of sales and service
(potential) clients as well as their opportunities
employees.  Insurers could lose the ability to cross-
Organisations wishing to exploit Super sell and up-sell if no direct contact can
Apps have to decide whether they want to be made with the clients
launch their own Super App along with  Establishing one’s own Super App
suitable partner services or integrate their might need a high investment and
proprietary digital services in a third- strong brand proposal
parties Super App marketplace.

56
Super Apps
Assess
Connected Experience

WeChat and Grab are just two examples of well- Share of consumers In October 2022 Elon Musk
interested in a Super twittered about “(…)
established Super Apps in Asia
App by country
More than a decade ago WeChat started its journey as a social platform, soon creating X, the everything
enhancing services like payment and shopping features. Today, the estimated U.S.
app”, which can be seen
number of monthly users is more than one billion in China alone. Germany
U.K.
as an announcement for
Grab is famous for food delivery, leading Southeast Asia’s market in this area. The app
also features transport and postal services, along with insurance services such as ride
Australia planning the creation of a
and travel policies. 0% 20% 40% 60% 80%
new Super App.
In 2021 it was reported that they were selling 100 million policies in less than two years.
Very or extremely interested
Slightly or somewhat interested

Key Insight

Insurers need to watch the evolution of this

trend closely. The ability of integrating suitable
white labelled or branded insurance services in
Super Apps might become a “must have” for
the accessibility of a wide range of mobile
customers in future.

57
Ubiquitous Connectivity
Assess
Connected Experience

The ability to connect to the internet from virtually anywhere

underpins many of our day-to-day activities. Ubiquitous
connectivity describes a state where devices can create,
share and process data — and never lose connectivity.

The concept of ubiquitous Opportunities

connectivity connects devices at Increased connectivity allows
any given location, regardless of companies to provide more
the service provider. This valuable services and
approach can bring advanced experiences through the use
technology to under-served of more intelligent and
regions for example via Low interconnected systems. How is it being used?
Earth Orbit (LEO) satellites and From self-driving cars to
thus enable companies to offer
Risks
facial-recognition systems,
more valuable products and As the number of connected the potential uses for
devices rises, the likelihood of ubiquitous connectivity
services through the continuous
data breaches and disruptions are almost limitless.
flow of data. Overall, ubiquitous
to business operations also
connectivity holds the potential increases, causing concern for
to significantly impact the way individuals, companies, and
we live, work and interact as governments regarding
a society. data privacy.

58
Ubiquitous Connectivity
Assess
Connected Experience

Low-power wide-area network WI-FI 6 The number of

(LPWAN) is a developing technology that is ideal By using Wi-Fi in the 6 GHz band, IoT technology 5G connections
for connecting devices that have low data usage can achieve significantly higher bandwidths. The is estimated to
and low transmission rates over large geographical higher the communication speeds of a network of increase to around
areas. This makes it an optimal option for IoT devices, the more dependable the system will be. 4.4 billion worldwide
devices that connect with each other through Wi-Fi 6 is an excellent choice for enhancing the
by 2027.
machine-to-machine interactions. LPWANs performance of smart home IoT networks.
are more energy-efficient, and therefore more
cost-effective.

Key Insight

Research forecasts that subscribers using Low Earth One of the key obstacles that IoT networks have encountered
Orbit (LEO) satellite services will reach 2.4 million is the limitation in wireless data transfer speeds.
in 2023, and the global number of LEO satellites As these technologies advance, so will the capabilities of IoT
approved for deployment will reach more than devices, such as sensors, edge computing, wearables, and smart
30,000 by the end of the decade. homes. There has been a significant increase in the deployment
of newer connectivity options, such as 5G, Wi-Fi 6, LPWAN and
ABI Research satellite connections, which make IoT solutions more viable.
PR Newswire 2023
59
Our Solutions
Connected Experience
Munich Re
apinity
Smart marketplace for insurance APIs
APIs are increasingly becoming a fundamental part of business strategies in the
insurance industry. Companies that now rely on API marketplaces, which integrate
seamlessly into the existing system landscape and support all existing APIs, will secure
the unfair advantage for the future.
The out-of-the-box solution of apinity allows insurance companies of all kinds to start to
consume, publish, and offer their APIs in a ground-breaking, new way. Are you ready to
boost your insurance business?

Our customers and business partners have often invested a lot of

money and time in their cloud and API management projects, but
were still far from being able to use their APIs efficiently in
daily business. We want to support customers where they
can benefit most from APIs on a daily basis: in the
intelligent use and leveraging of existing data.

Anna-Carina Häusler

Managing Director apinity

60
Our Solutions
Connected Experience
Munich Re Faster time to market Plug-and-play APIs
Realytix Zero Reduce time to market to just a Very strong API capabilities enable easy
Florian Next-level automated few weeks thanks to unique self- integration of multiple external data
Niklas underwriting solutions configuration capabilities and a sources or modules (such as AI or machine
Head of Realytix no-code/low-code platform to learning algorithms) at any point in the
Realytix Zero is a global, scalable
implement and adapt products underwriting process - with a plug-
technology platform for primary
very quickly. and-play approach.
insurers, brokers and MGAs that want
to digitalise and automate the
underwriting of single-risk business.
As of today, we still see so Realytix Zero is set up as a no-code and
many manual and non- Software as a Service platform offering
UI and API capabilities and therefore
digitised underwriting opens up entirely new opportunities in
workflows in many lines of the development and distribution of
business. This can be non-life insurance products.
No development Munich Re expertise
changed immediately with Its unique self-configuration capabilities
and vast API integration possibilities No need for in-house development – Access to Munich Re’s analytics power
Realytix Zero in a couple of allow rapid implementation and just use our proven platform. and decades of pricing and underwriting
experience.
weeks only. adaptation of products.

Benefits at a glance
Let´s start making the
 Significantly reduced time to market
underwriting world more  Customisable, flexibly implementable
digital and increase your platform
digital distribution channels  Digital distribution
 Increased (process) efficiency,
– reach out to us. cost and time savings

61
Our Solutions
Connected Experience
Munich Re Munich Re
Cert2go One Cat
cert2go – speed up your Comprehensive and rapid response
(re)insuring process for to natural catastrophes
single risks When a devastating natural catastrophe
cert2go is a cloud-based underwriting strikes, companies and organisations
platform for clients seeking rapid, efficient need their insurance claims to be settled
and automated capacity from Munich Re. quickly, in order to avoid cash flow
problems. Traditional covers often only
Our target group consists of clients who
respond after lengthy claims adjustments
turn to Munich Re in search of cover for
and leave gaps in costs associated with
their small and medium-sized single risks.
restoring to the pre-event level. To close
The platform provides clients with these gaps, Munich Re has designed “One
customisable solutions for many lines Cat”, a tailor-made solution concept with
of business and markets and covers the an unprecedented level of transparency
whole underwriting process. The core and a very simple payout process.
of the platform is Munich Re’s pricing
engine, based on decades of Benefits at a glance
underwriting expertise.  Parametric triggers ensure
rapid recovery
Benefits at a glance
 Covers previously uninsurable risks
 Accelerates underwriting processes
arising from natural catastrophes
 Minimises administrative workload
 Unprecedented level of transparency
 Fully digitalises the
underwriting process  No deductibles
 Customisable for different lines of  Reduced claims-related expenses
business and markets

62
Data & AI
Technology impacts in 2023

54%
During 2022

The super app 37% of organisation leaders

Grab sold of publicized automotive
believe AI-driven decisions
100m IoT exploits were
conducted by friendly
Cyber crime is
are interpretable and
policies in less going to cost the world
researchers. easily explainable.
than 2 years.
US$ 8tn
in 2023. The market for
Digital twin
digital twins is
technologies can improve
product quality by up to
68% growing to

US$ 6.69bn
of data in organization
25%. is not
leveraged.
in manufacturing
by 2025.

64
All trends 2023
Data & AI
The trend field of Data & AI is the home for data-driven technologies and enterprise solutions based What is new in 2023 in the field of Data & AI? Large language models are emerging and have the
upon artificial intelligence and data. A key technology in this field is, for instance, natural-language potential to impact various industries, including insurance. They may form the basis for end-to-end
processing which enables machines to interpret human speech. Machine learning becomes an automatization solutions, but also yield challenges due to immaturity and compliance questions.
enabler for processing the data and making autonomous decisions. Reinforcement learning and With decentralised data architectures as well as data mesh technologies, the industry might be able
knowledge graphs are only two principles that help machines become artificially intelligent. to organize and exploit data in new ways. Great potential lies in combining the data and AI
techniques described in this trend field.

Adopt – Start initiatives in your unit Trial – Initiatives in affected units Assess – Evaluation needed Hold – Watch list
AI Democratisation Data Mesh Multimodal AI Artificial General Intelligence
AI democratisation refers to making AI Data mesh refers to a decentralised data Multimodal AI combines inputs from different It may be possible to build a
technology available to the wider public. architecture that organizes data by a specific sources of data to achieve more accurate machine with human cognitive capabilities –
business unit. results. but we are probably decades away
from developing it.
Digital Twin Knowledge Graphs Reinforcement Learning
A virtual counterpart of a real object that enables Knowledge graphs are used to enhance search Reinforcement learning (RL) is an iterative
IT systems to interact with it rather than the real results by creating relationships between data learning process where a software agent faces a
object directly. from different sources. game-like situation and employs trial and error
to find a solution.

Machine-Driven Decisions Large Language Models

Business decisions that are derived and LLMs gained significant attention in recent
backed by verifiable, quantitative data analysis. years for the ability to generate human-like
text, images and videos.

Natural Language Processing

Natural-language processing (NLP) can ease Quantum Computing
human-computer interaction and leads to Quantum computing stands for unprecedented
machines understanding and acting on text. computing power. It could trigger a new wave of
technological development over the next five to
ten years.
Trustworthy AI
It leads back to a human ethic discussion based
on how society, economy and politics aim to
guide the use of AI technology. new in 2023

65
Trend Evolution
Data & AI

Evolution over time

Adopt Trial Assess Hold

Start initiatives in your unit Initiatives in affected units Evaluation needed Watch list

AI Democratisation 2023 2022

Artificial General Intelligence 2023 2022 2021 2020 2019

Data Mesh 2023

Digital Twin 2023 2022 2021 2020 2019

Knowledge Graphs 2023 2022 2021 2020 2019

Large Language Models 2023

Machine-Driven Decisions 2023 2022 2021 2020 2019

Multimodal AI 2022 2023

66
Trend Evolution
Data & AI

Evolution over time

Adopt Trial Assess Hold

Start initiatives in your unit Initiatives in affected units Evaluation needed Watch list

NLP 2023 2022 2021 2020

Quantum Computing 2023 2022 2021 2019 2020

Reinforcement Learning 2022 2023 2021

Trustworthy AI 2023 2022

67
Impacts on the Value Chain
Data & AI

Explanation Insurance Value Chain

The illustration represents the impact the trend
fields have upon on the insurance value chain. Product Sales & Risk Customer Asset
With the dark boxes reflecting areas of high Design & Pricing Distribution Underwriting Policy Admin Management Engagement Claims Management
impact, and light boxes low impact.

AI Democratisation

Artificial General Intelligence

Data Mesh

Digital Twin

Knowledge Graphs

Large Language Models

Machine-Driven Decisions

Multimodal AI

68
Impacts on the Value Chain
Data & AI

Explanation Insurance Value Chain

The illustration represents the impact the trend
fields have upon on the insurance value chain. Product Sales & Risk Customer Asset
With the dark boxes reflecting areas of high Design & Pricing Distribution Underwriting Policy Admin Management Engagement Claims Management
impact, and light boxes low impact.

NLP

Quantum Computing

Reinforcement Learning

Trustworthy AI

69
AI Democratisation
Adopt
Data & AI

AI democratisation refers Likewise, low-code or no-code  Low-code/no-code tools present

ways to overcome organically grown
platforms enable people with limited
to the development of programming knowledge to develop complexity of documents and processes
that end up in an “excel hell” that is
tools that make AI
applications using point-and-click, or
other graphical interfaces, and model- impossible to maintain

technology available driven logic. These advances enable  Prototypes and minimum viable
products for applications are produced
AI democratisation, giving experts
to all companies from different industries (e.g., healthcare, in higher quality and in a timely manner
that increases their business value
manufacturing, agriculture) the power to
and individuals. use AI for their respective needs.  Any employee can become a “citizen
developer” and take ownership for their
Previously, AI was usable only by a The move toward AI Democratisation application with reduced IT overhead
select few data scientists, analysts and enables a bigger group of professionals
engineers. However, with the advent to make sense of and exploit vast Risks
of features such as AutoML and Low- amounts of data. This step can improve
 The challenge for IT and business is to
Code/No-Code, the once today’s healthcare processes and incorporate quick and easy prototypes
mysterious world of AI is increasingly results. More insights on patient care, into sustainable IT artefacts
becoming used by non-data scientists. for instance, can improve predictive care
 Low code and no code platforms are
and early diagnosis of individuals. limited to simple applications. They are
With AutoML, the process of machine
not suitable for supporting complex
learning is shortened so that the user business solutions
only has to load the data.
 System legacy debt limits the full
The automated process then allows
potential of low code or no code
the input to be interpreted without
the need for the programming-
Opportunities platforms, especially in terms of data
integration
intensive middle layer, as is the case  Entry to the world of AI and the
possibility to apply its potential in the  Enterprise-specific compliance
with traditional machine learning. guidelines and data privacy standards
insurance sector without the need for
deep expertise in programming will need to be in place

70
AI Democratisation
Adopt
Data & AI

Dataiku 11 – Help business Benefits of low-code development

users to embrace AI
Here in Global Consulting, the Source
consulting arm of Munich Re, Dataiku Scalability Capability
we work closely with our An Enhance Toolset to Scale AI
primary insurance clients Dataiku 11 is an update to the company's
helping them grow profitably. data science and AI platform. The goal
of this product is to make AI accessible
Quite recently, we leveraged to all businesses. This is achieved with
the full capabilities of fore-casting, prescriptive, coding and
AutoML to help a digital deep learning tools. Agility
insurer estimate future
costs and customer Pricing
behavior. The result was Pricing and underwriting represent
Key Insight
reflected in a projected Loss a key pillar of primary insurance.
Ratio improvement of 4%. Developments in technology and
AutoML significantly shortens the time to develop and test new
digitalisation have boosted the
finesse of pricing to the next level.
models. It allows less machine-learning experienced employees to
select the right models and initial parametrisation. However, these
The recent emergence of the
often SaaS-based tools still require interpretation and testing before
automated machine learning field is
Dr. Massimo changing the pricing department’s being deployed in a production system. Depending on the use case,
Cavadini role, bringing it to the center of the additional work by data scientists may still be required.
decision-making process and having a
Senior Executive Partner
huge impact upon the governance of
insurance pricing teams.
71
Digital Twin
Adopt
Data & AI

A digital Twin is a virtual Opportunities

counterpart of a real  Digital twins allow exploration and
improvement in production processes
object enabling IT without expensive physical prototypes,
systems to interact with which will be particularly important
in the era of the Internet of Things (IoT)
it directly, rather than  Digital twins promise to allow for
with the real object. shorter development cycles and
massively increased options to test
A digital twin helps to improve digital prototypes much better, so that
maintenance, upgrades, repairs and the real-world product will be less likely
operation of the actual object. For to malfunction
example, it could be a model of a
sound system enabling a remote Risks
user to control the physical system
 Particularly for products with long
with buttons on a mobile device.
lifetimes, such as buildings, aircraft,
Digital twins can also be used for ships, etc., digital twins might become
product development, as they enable unusable at some point (due to
product testing and simulations software issues)
without the need for constructing a
 Firms employing digital twins provided
physical object, thereby driving
by third parties face a lock-in risk with
innovation efficiency.
their vendor, since switching to other
Even though the idea of the digital vendors may become very costly at
twin is still evolving, hundreds of some point
millions of things will most likely have  A digital twin could be hacked in order
digital twins in the future. to fake damage which, in turn, triggers
a claim payment from the insurer

72
Digital Twin
Adopt
Data & AI

Earth's Digital Twin Data from Smart Vessels HSB and HELIXintel
Industry Industry Simplifying equipment
Various Marine maintenance with digital twins The HELIXintel platform encourages
facility managers to take better care
Source Source HSB has teamed up with HELIXintel to of their property & equipment through
Nvidia Siemens.com offer an innovative software solution to preventative maintenance thus
help Equipment Breakdown policy- maximising useful life and reducing repair
Climate research Digital twin of a ship costs. The result is tremendous savings
holders and Sensor Solutions by HSB
Lockheed Martin and Nvidia are planning The ship’s digital twin connects to for the business and a better risk for
customers thrive in today’s tough
to build an AI-driven Earth Observation onboard sensors to gather real-time HSB and our client partners.
economic environment.
Digital Twin. The aim is to monitor current operational data to be used for various
HELIXintel has created an ecosystem Rebecca Zicko
global environmental conditions, such as purposes (e.g., from planning
extreme weather events. maintenance to optimising fuel that includes building owners, facility Equipment Breakdown
use and route planning). managers, service providers, OEMs and Strategy VP, HSB
utility companies where they can work
together to utilise data integration, and
The market for digital twins is growing rapidly in a range of machine learning in a way that benefits
sectors all entities. The HELIXintel platform
Key Insight
Market value in billion U.S. dollars (World Economic Forum, 2022) creates a “digital twin” of the property,
facility and equipment inventory. These
6,69
records are not only highly organised and
5,06 5,09 Digital representations of real-
readily accessible, but critical data is also
3,83 3,81 sourced and matched to assets - to world use cases allow for much
2,26
2,83 enable data-driven property cheaper planning, testing and
management. optimising. As the adoption of
0,59 0,49 0,43 0,34 0,29 0,23 0,28 digital twins grows, so do
Manu- Automotive Aviation Energy Healthcare Logistics Others insurance-related initiatives
facturing and utilities and retail and partnerships.
2020 2025
73
Machine-driven Decisions
Adopt
Data & AI

Business decisions that big data analytics are used in every Opportunities Risks
sector of the world economy.
are derived and backed Recognised as a global trend, this  High-quality statistical  In order to implement machine-
by verifiable, quantitative growth in available data is enabling predictions through machine- driven decisions, some decision
driven decisions could free criteria need to be implemented at
data analysis and executed valuable insights and assessments
employees from repetitive tasks the beginning, and on-the-fly
to be made.
machines with some or no  Unlike individuals, machine-
adjustments are virtually
Today, algorithms scan every impossible. Therefore, wrong
human involvement. driven decisions stick to clearly
decisions at the outset are likely to
element of data gathered about a defined decision criteria and are
become expensive over the course
A tremendous increase in data specific issue – such as the field not influenced by personal moods
of the project
has contributed to the rise of a of interests of a certain client – and that might negatively affect
extract all relevant information. decision-making  Distortions or changes in the
“data-driven” era, where
Conclusions are derived and logical informational content of the data
 High volume, real-time decision-
decisions made based on this rich underlying the decision algorithm
making like fraud checking at
set of information. Depending on might result in unintentional and,
scale is supported by AI and
in some cases, even
the context, these decisions may be cannot be carried out by humans
systematic errors
made without any human
 Sensors can automatically detect
involvement.  Unfairness in machine-driven
danger, safeguard humans and
decision-making due to bias during
Given the autonomy of these initiate interruption so that faulty
data gathering or implementation is
machines, the success of a decision machines do not break
a risk that must be avoided
completely
not only relies on the quality of
the data harvested, but significantly  Machine-driven decisions can
also on the sophistication, quality, also automate claims payments in
and fairness of the pre-determined parametric insurance scenarios –
analysis and interpretation once the required proof of loss
parameters. has been provided by the insured
74
Machine-driven Decisions (1/2)
Adopt
Data & AI

Human vs. algorithmic Selected application fields of algorithmic

decision-makers decision-making
Source Aircraft collision avoidance
Wesche et al., 2022
Distributed wildfire surveillance Sensor-equipped
People’s reactions to decisions by human vs. drones monitor
wildfires. This
algorithmic decision-makers Automated driving
requires reasoning
With new AI applications available to organisations, various processes about the stochastic
that were previously carried out by humans can now be replaced with Breast cancer screening evolution of the fire.
automated decision-making. This new decision-making power of
algorithms, however, raises concerns regarding trust, acceptance, Mars science exploration Rover discoveries are
and perceived transparency. impaired by weak
communication links to
Wesche et al. conducted tests in realistic contexts to study the Financial consumption and
operation teams on Earth
impact of human vs. algorithmic decision-makers. They found portfolio allocation (an exchange takes
that algorithmic decision-making in personnel selection can 30 minutes per direction).
negatively impact trust and acceptance. Companies should well Greater levels of
consider which decision-making tasks they delegate and try to autonomy can make
meet employees’ informational needs. science exploration
missions 5x more efficient.

75
Machine-driven Decisions (2/2)
Adopt
Data & AI

Three levels of machine- Is AI ready to make unsupervised decisions? Key Insight

driven decision-making Recent incidents have raised doubts about AI as an ultimate
Higher-level decision support, decision-maker, e.g.
in which decisions are primarily
made by humans, “based on  Recruiting bias due to tainted data
principles and ethics, experience  Self-learning bots learning offensive language and incorrect facts
and bias, logic and reasoning,
emotion, skills and style.”  Bad advice from experimental healthcare chatbots Trends towards
automation and
How can leaders assure greater humanity as these systems
proliferate?
autonomous things
Augmented machine support,
in which machines and AI are enabled by
 Encourage and build an organisational culture and training
“generate recommendations,
that promotes ethics in AI decisions systems that draw
provide diagnostic analytics
for human validation and  Remove bias from data
conclusions and take
exploration.” action based on the
 Keep humans in the loop
data they receive.
 Validate before deployment in real-world scenarios
Highly automated settings,
While machine-driven
in which there is still a need  Teach machines human values decisions can speed
for “guard rails or a human-
in-the-loop for exceptional
up processes, their
cases.” emergence also
raises concerns.

76
Natural-language Processing
Adopt
Data & AI
Opportunities
Natural-language processing (NLP) can ease  Automate repetitive tasks for efficiency
human-computer interaction and leads to and cost savings
 Improve customer service with chatbots
machines understanding and acting on text. and virtual assistants
 Make better decisions by analysing large
Natural Language Processing (NLP) is the study of the interaction between amounts of unstructured data
computers and human language. NLP enables computers to process, interpret,  Detect fraud by analysing text data
and respond to natural language data and includes tasks such as machine  Personalise products and services
translation, text generation, sentiment analysis, and information extraction. for customers
 Improve risk management by extracting
The field has progressed dramatically in recent years. The advent of large structured data from unstructured data
language models like GPT-3 and ChatGPT has especially sparked the
imagination of what is possible with NLP (see section “Large Language
Models”). While such large language models are very resource hungry or even
still in experimental stage, language models in general have already helped Risks
enterprise solutions to substantially reduce the amount of training data  Bias in data and models, leading to
required to e.g., extract meaningful information from text with high accuracy. unfair or inaccurate results
 Dependence upon large amounts of
In addition, due to data privacy regulations, there are limitations on the usage data, which may be scarce or hard
of external services like ChatGPT for processing sensitive information such as to obtain
insurance contracts, which can impact the potential applications.  Difficulty in explaining the model's
decisions and lack of transparency
For such situations, a range of compact pre-trained NLP models is available.  Reliance on proprietary systems, making
While they are not as mighty, they can be hosted internally and fine-tuned to it hard to adapt to new technologies
meet the specific demands of a use case. or vendors
 Fake data and misinformation, leading
to wrong conclusions and decisions
based on false information
77
Natural-language Processing
Adopt
Data & AI

Medical Pre-assessment service Examples for application of NLP in Insurance

NLP The pre-assessment service allows insurance
carriers to calculate non-binding quotes from
 Improve risk assessments by analysing policies and loss reports
 Accelerate claims processing by automated information extraction
broker emails. A large set of NLP models is
Industry  Enhance underwriting models by incorporating media information
extracting key demographic and medical
Insurance
information from the unstructured text. The pre-  More efficient client interaction by automating aspects of communication
assessment service can help insurance carriers to
Source
quickly and accurately provide quotes to potential
Munich Re
customers, which can help to streamline the sales
process and increase the chances of closing a sale.
Current Trends in NLP
 Large Language Models and ways to make them better applicable to use
cases and more efficient even with less parameters
 Low-resource NLP: Developing NLP techniques that can still work well even
when there is little data to work with
 Multimodal NLP: Incorporating other types of information, like images and
videos, to help computers understand what people are saying
 Adversarial training: Teaching computers to recognise and defend against
While impressive breakthroughs in large language models like ChatGPT showcase
the power of NLP, the practical application of NLP in an enterprise setting currently attempts to manipulate them
often requires models that can be operated in-house for cost and data privacy  Explainable AI: Designing NLP models to be more transparent, so that
reasons. Recent advancements in language models have allowed companies to
create solutions with little training data. people can better understand how the computer came to a certain
Dr. Matthias Kaper
conclusion

Senior Data Scientist

Data and Analytics
78
Trustworthy AI
Adopt
Data & AI

Artificial intelligence (AI) AI systems, we see high sensitivity  Trustworthy AI could also help with the
around bias in data, discrimination and proper use of data and exclude
has reached a new level of fairness in the US. In contrast, China unnecessary data from the decision-
public awareness with the has just started working on its first set making process. This includes using
of guidelines while Singapore, for data from informal inquiries for
launch of openAI’s ChatGPT. improper identification of fraud
example, introduced the world’s first AI
All studies expect Governance Testing Framework and
exponential growth for AI. Toolkit in 2022 for companies who want Risks
to demonstrate responsible AI in an
Despite this prosperous objective and verifiable manner.
 There is a vast number of attempts for
AI governance. International
outlook, AI triggered a broad standardisation efforts are ongoing, but
discussion about how not yet offering comprehensive
applicable AI frameworks
society, economy and politics
 Lately, in Europe, AI in life and health
aim to guide the use of AI insurance is classified as high-risk
technology and make it Opportunities
requiring extensive validation and
verification
trustworthy.  Low/no-code AI models as well as
 Trustworthy AI will become a
competitive edge for companies. publicly available large AI models
Scalability of AI requires humans to Stakeholders (e.g., customers, media, (e.g., ChatGPT) will challenge all efforts
trust the safety, accuracy, fairness and financial analysts, rating agencies, regarding potential explainability of
integrity of outputs and decision- new talents) have an increasingly AI-based results
making processes. Different regions, strong interest in how AI is developed,
societies, cultures as well as legal and deployed and operated
regulatory backgrounds have different
approaches to this challenge.  Implementing trustworthy AI will
ensure objectivity, fairness and non-
While Europe prefers a rather strict discrimination in decision-making
risk-based classification across all related to pricing, claim processing
and approval

79
Trustworthy AI (1/2)
Adopt
Data & AI

European AI Act
As proposed by the European Commission, 2022
Each AI system will be classified in one out of 4 risk categories requiring verification and
validation of such systems.

Unacceptable Risk

High Risk

Limited Risk (AI systems with specific transparency obligations)

Minimal Risk

Step 1 Step 2 Step 3 Step 4

Trustworthy AI will
A high-risk AI system It needs to undergo the Registration of stand- A declaration of If substantial changes become the driver for
is developed. conformity assessment alone AI systems in an conformity needs to be happen in the AI
and comply with AI EU database. signed and the AI system’s lifecycle scaling AI technology
requirements. system should bear the
For some systems a CE marking. The
notified body is system can be placed
involved too. on the market. Dr. Oliver Maguhn
Senior Project Manager AI
Go back to Step 2 Co-Founder CertAI
80
Trustworthy AI (2/2)
Adopt
Data & AI

Implementations for responsible AI in organisations Key Insight

Planned actions by companies in the United States (selection)

Confirm AI is compliant with

applicable regulations
40% 57%
Confirm AI-driven decisions are
32% 54%
The need for
interpretable and easily explainable
trustworthy AI is
Develop and report on controls
related to AI models, processes
31% 48% recognised by industry
leaders, policymakers,
Address the issues of fairness 35% 46% and solutions providers.
Review to be sure third-party However,
services meet standards
27% 41%
implementation will
No plans to address AI
1% differ across regions.
responsibility issues

Others Leaders

81
Data Mesh
Trial
Data & AI

As companies seek to leverage For instance, data product owners are Risks
supposed to provide data catalogues
the vast amounts of data they  Without proper planning and
with descriptions and interpretation
accumulate, centralised instructions, thereby enhancing the governance, a data mesh can easily
repositories like data warehouses usability of data sets. gain complexity, driving up costs
and data lakes encounter  For data mesh architectures to be
scalability issues. A so-called successful, data nodes must
be discoverable
data mesh, on the other hand, Opportunities
 Computing very intensive operations
aims to eliminate problems of  Data in centralised repositories may in a data mesh environment can be
data availability and accessibility, struggle with lack of transparency. A time-consuming
thereby democratising data and data mesh approach can provide
 Companies employing a data mesh
structure and encourage purposeful
fostering innovation. need to find the right balance between
usage of data
decentralised data ownership and
A data mesh refers to a decentralised data
 Viewing data as products with a ‘job centralised data governance
architecture that clusters data by corporate
to be done’ promotes value-based
functions. This distributed architecture is  Adopting data mesh requires cultural
thinking throughout a company
advantageous in terms of data request and organisational change
scalability.  Data mesh may foster innovation at
scale by making data accessible to a
While data tends to aimlessly ‘swim’ in
distributed workforce while enhancing
data lakes, a data mesh allocates a
the user experience
dedicated data product owner to each data
set. The data product owner is tasked with  A data mesh approach enables
curating, offering and marketing their data business units to define data products
to users across the company via a shared that render business value
data mesh platform. In addition to  A data mash project can start small
decentralised data ownership and and grow as demand grows
accessibility, the data mesh approach
highlights the role of user experience.
82
Data Mesh (1/2)
Trial
Data & AI

The four principles of data mesh

Domain ownership Data as a product Self-service data Federated computational

Domain ownership reduces the Applying design thinking for and analytics governance
hops between analytical data data assets. Encapsulating Removing friction and Automating data governance
consumers and data sources. related code, policies and technological complexities from policies without a centralised
infrastructure in a cohesive the interaction between data authority.
product. producers and consumers.

Zhamak Dehghani
director of emerging technologies at Thoughtworks North America, is

68%
often referred to as the inventor of the data mesh approach.
In 2019, Deghani introduced the concept in her article ‘How to Move
Beyond a Monolithic Data Lake to a Distributed Data Mesh’.

of data in organisations is not analysed

83
Data Mesh (2/2)
Trial
Data & AI

Data Mesh Architecture Key Insight

Interoperability Documentation Security Privacy Compliance Governance

Policy Policy Policy Policy Policy Group

In many cases,
Enabling
Domain Team centralised data
Domain
Domain Analytics Team platform architectures
Consulting cannot deliver insights
Data
Domain with the speed and
Example
Product
Operational Data Data
flexibility that is
Data Product Contract needed. A data mesh
Best Practices
architecture can
enable users to
collect, sort, analyse
Storage and Access Policy
Data and display
Catalog Monitoring Platform
Query Engine Management Automation
Team data in a quick and
understandable way.

84
Knowledge Graphs
Trial
Data & AI

Knowledge graphs are used to enhance search Opportunities

results and deliver valuable insights by creating  Graph-based solutions enrich data
relationships between data from different sources. with missing data. Their capabilities
can identify entities, such as people,
AI applications in companies cannot reach their full potential when data by analysing unstructured data from
is treated in silos and not fused with reasoning. Knowledge graphs various sources
are recognised as a new wave of AI that brings together machine learning
and graph technologies to overcome these obstacles. The goal is for AI to
 By building up rich knowledge Risks
networks from unstructured data,
gain context and to process and store information close to the way a graph-based technology can  The fate of knowledge graphs is closely
human brain does. Humans constantly interpret information using background compensate for the lack of human tied with the platforms in which they
knowledge and common-sense. A knowledge graph similarly represents reasoning in AI solutions are implemented
semantics by describing entities and their relationships in a graph-  With their contextualising capabilities,  Designing, maintaining and exploiting
like structure. This process is called semantic enrichment and utilises natural- knowledge graphs draw connections knowledge graphs requires strategic
language processing (NLP). The result is a complex network of between data that would otherwise intent and data management skills.
information from which missing facts can be inferred far better than from remain distinct. They enable advanced For example, to semi-automate the
a conventional database alone. The technology has valuable implications for pattern recognition and anomalous construction of such graphs or determine
behavior detection for use in fraud which open data repositories should be
companies. Prominent use cases are search engines, question-answering
detection, for example combined with enterprise data
services and social networks. But any company faced with disconnected,
heterogeneous and ever-growing data sources can use knowledge graphs to  Knowledge graphs do not necessarily  In addition, governance, operation
avoid data overload and generate big data value. Knowledge graphs facilitate come as stand-alone AI market and maintenance guidelines should
solutions. Their potential lies in be put in place
queries and help integrate company data with other internal as well as
implementation in broader AI and
external data, yielding novel insights.  Companies need to employ methods to
cloud platforms and services
retain semantic coherence across
the graph

85
Knowledge Graphs
Trial
Data & AI

AuraDS Tiger Graph DB

Industry
Industry Various
Various
Source
Source Tiger Graph
Neo4j
Distributed Native Graph
Graph Data Science as a Database as a Service
Tiger Graph Cloud introduced a tool that
Service
allows easy access to knowledge graphs
Neo4j is a company that enables without any extra hardware.
organisations to build new applications
that leverages connections in data
with the help of knowledge graphs.
Patents4Science
AuraDS is a fully-managed graph
analytics and machine learning engine. Industry
It connects billions of data points Intellectual Property
In order to properly understand and analytically use data to identify connections and generates
a visualisation. Source
in a powerful way, it is important to semantically connect
This product can be used in anomaly FIZ Karlsruhe
data across multiple platforms and domains.
and fraud detection, customer 360,
Analysis of Patent Data
and network analysis.
Patents4Science uses knowledge
Latifa Mahna graphs, NLP and other technology to
search for and connect data.
Consultant Data Engineering,
Data and Analytics

86
Large Language Models
Trial
Data & AI

Revolutionizing Thus, LLMs increase the feasibility of

insurance-specific solutions since less
 Human-AI collaboration using
instructions from natural language
insurance with Large manual input, such as in the form of text
annotations, is needed from domain Risks
Language Models experts. Consequently, next generation
LLMs will allow us to use NLP  Unfair treatment of certain groups of
Large language models (LLMs), such as customers, such as those from minority
methodologies in areas where this was
GPT-3/chatGPT, PaLM, and Claude, backgrounds
previously not economic and may form
have gained significant attention in
the basis for end-to-end automatization  Difficulty in ensuring the quality of the
recent years for their ability to generate
solutions in claims handling or data used to train the models and the
human-like text and perform a wide
underwriting, for example. potential for errors to propagate
range of natural language processing
tasks. These models have the potential  Models may violate the copyright or leak
to revolutionize a number of industries, confidential or private information
including insurance. contained in its training data
LLMs are deep learning models with  Through the use of LLMs, the
millions, even billions of parameters Opportunities interpretability and transparency in the
trained on terabytes of text generating decision-making process may suffer
 Automation and improvement of
probability distributions over sequences  LLMs may hallucinate, that is generate
insurance-related natural language
of words, for example, to predict the tasks text that is not grounded on real existing
next word in a sentence. They are events, objects or facts
further fine-tuned to align with human  Extracting new risk sights from
instructions, for example, via unstructured data
reinforcement learning from  Natural Language Interfaces to make
human feedback. complex insurance products more
Another benefit of LLMs is that they accessible
require little or even no additional data  Improving customer engagement and
to adapt them to new NLP tasks experience through natural language
(“zero/few-shot learning”). chatbots and virtual assistants
87
Large Language Models
Trial
Data & AI

Platforms AI applications are on the edge of Leaps in research A paradigm shift in

revolutionising industries using LLMs Two aspects accelerate the advancements of research state-of-the-art
on new A wave of recent breakthroughs in LLMs sparks the and applicability of Large Language Models models is taking place
from an increasing
horizons potential to reshape many domains entirely and pave  Emergence: With the increasing number of
new ways for people to interact with technology. parameters in the LLMs, the models can number of parameters
perform new tasks not specifically trained for. to ingestion of domain
Leading tech companies, startups and research-
Industry Examples are logical reasoning or solving data and more
organisations race to develop the most versatile and
multi-step math problems complex architectures.
Various capable LLMs that serve AI applications as a platform,
 Homogenisation: A rising breadth of data While LLM
impacting billions of users.
trained into a single model covers an increasing capabilities are
Source In an ecosystem-partnerships approach organisations variety of domains and applications
join forces to explore new use-cases, share development growing exponentially,
Munich Re
costs and risks. there is a risk faults or
inappropriate data
built into the models
LLM Capabilities may get distributed
Crunchbase Key factors for Curve instantaneously to the
State of AI user mass.
industrial adoption
Enterprise software is the most Accessibility, safety
invested category globally and trustworthiness Andreas Schumacher
Organisations and start-ups in the AI NLP Access to leading LLMs is limited and Project Manager Artificial Intelligence,
field achieve tremendous valuations and owned by a few companies holding
Impacts on the insurance value chain Data and Analytics
venture capital, such as Anthropic, raising proprietary IP and the required The ability of Large Language Models to process and understand
$740 million within one year of being infrastructure. Offered as a service, natural language opens groundbreaking new possibilities across
founded. OpenAI, after its recent the use is delineated to the API- industries and redefines interaction with technology. In insurance
response of the models, while providers they propel the capacity to unlock the full potential of data,
achievements, received an additional
keep architectures, training data and automize complex processes, drive combinatorial innovations with,
$10 billion from Microsoft. parameters behind closed doors.
e.g., the internet of things and provide personalized experiences
that were once unimaginable.”
88
Quantum Computing
Trial
Data & AI

The unprecedented power Opportunities

of Quantum Computing  QC holds the promise to solve open problems in long term weather and climate
predictions, battery development, finance, pandemics, logistics, cyber and AI
could trigger a new wave  The success in sampling from distributions may point to early commercial
of technological applications in estimating extreme risks or asset pricing using Monte Carlo methods
advancement over the  There is high probability of substantial advancement in QC technology over the next
few years due to the availability of many different technological paths. This should
next five to ten years. result in more business relevant use cases
 New technologies are creating new risks, some of which can be insured while some
Quantum Computing (QC) holds the need to be tackled in a different way
promise to solve many problems, that  Due to the global QC race, there are many initiatives that should drive research and
currently cannot be solved with the development of quantum solutions
traditional computing.

It's currently based on technologies, like superconducting elements, trapped ions or

photons or other more exotic technologies. Sampling from distributions is a first use Risks
case, where QC supremacy was claimed by Google (2019), USTC (2020) and  The speed of future development and especially the timing of quantum advantage for
Xanadu (2022). different use cases is not easy to predict
Quantum computers use so-called qubits, which can represent 0 and 1 at the same  QC may endanger technologies such as encryption and Blockchain, whose security
time. Using a process called “entanglement” to combine qubits results in an exponential features rely on the limitations of current computers
capability to store and process information: Doubling the size of a 10-bit system will  QC will be based on many completely new algorithms. As with AI, many pitfalls and
double the performance, while it will increase the performance of a similar QC system constraints of QC will emerge during production time. Also required talent to
by a factor of 2^10=1024. produce/adapt these algorithms is scarce
In 2021, ten leading German companies, including Munich Re, set up the Quantum  The probabilistic nature of QC will provide challenges to many industrial applications,
Technology & Application Consortium (QUTAC) with the objective of accelerating the like interpretation, possibility of wrong results and missing audit trails
transition of QC from research labs to useable industrial applications. First results have
already been published.

89
Quantum Computing (1/2)
Trial
Data & AI

Cyber attacks Artificial Intelligence Weather Fertiliser Battery

QC will be able to break currently Artificial Intelligence (AI) and QC prediction development materials
used encryption algorithms. New are a perfect fit, as QC is potent but
encryption algorithms are planned currently noisy, while AI algorithms Industry Industry Industry
but hold and decrypt attacks need high computational power but Agriculture Chemistry Automotive
have already started. are noise tolerant.
Source Source Source
PASQAL IBM PsiQuantum
BASF collaborates Nitrogen fixation Mercedes-Benz
with PASQAL IBM predicts that in collaborates with
to predict weather five years, by utilising PsiQuantum
patterns. The AI and quantum to demonstrate
predictions generate computing, innovative how fault-tolerant
parameters that are solutions to enable
2022: Nobel prize for New algorithms used to simulate crop
quantum computing
nitrogen fixation at can accelerate battery
Quantum Computing QC achieves full potential when yields and growth a sustainable scale designs. The focus is on
quantum native algorithms are
The importance of QC was stages. The approach can be developed and quantum algorithms for
used. Time and expertise is
acknowledged by awarding the Nobel is to use quantum used to help feed the simulating effects of the
necessary to develop these for
prize to three scientists working on different business applications. neural networks to world’s rapidly electrolyte additive,
quantum information/ solve the underlying growing population. fluoroethylene
entanglement. differential equations. carbonate.

90
Quantum Computing (2/2)
Trial
Data & AI

First movers Key Insight

Industry
Financials
Source
HPCwire
JPMorgan Chase bets on Quantum Computing
Quantum Computing is coming
JPMorgan Chase believes that financial
services will be first to take advantage of and it will
quantum computing.
revolutionise the way
how we can understand
and simulate our world.
First applications will be
35.5bn Quantum Computing will redefine how complex
risks are estimated and computed and will open up
seen in the next few
years, but adaptation
a world of new applications. The technology will prove hard for
Public and private investments
is challenging, so we need to prepare now. companies, as
totalled US$ 35.5bn by 2022 across
a range of quantum technologies. algorithms and business
Dr. Andreas Nawroth models need to
World Economic Forum be changed.
Leading Expert Artificial Intelligence

91
Multimodal AI
Assess
Data & AI

Multimodal AI combines Opportunities

 The incorporation of various data
inputs from different types in the machine learning
sources of data to achieve process can improve performance
and lead to more accurate results
more accurate results. or insights
The sources of data can vary  The technology can also be used to
and include texts, audios and images, process unstructured documents
amongst others. By analysing these and alleviate work for employees
different types, AI could contextualise
the data and thereby achieve better
performance than traditional Risks
unimodal AI. Various industries  Multimodal AI is highly susceptible
including automotive, consumer to adopting biases in data sets,
devices and healthcare will, or could especially when there is a lack
be, disrupted by multimodal AI. of high-quality data with which
For instance, advanced machines can “learn”
driver assistance systems (ADAS)  With such powerful technology,
and human-machine Interface it could also generate equally
(HMI) assistants are adopting dangerous results if fed with
the technology. Consumer devices malicious pieces of data
are observing the use of multimodal
AI for security or payment
authentication purposes.
Healthcare is also exploring
this field in terms of diagnosis, by
incorporating medical images with
patients’ electronic health records.
92
Multimodal AI (1/2)
Assess
Data & AI

Can multimodal AI improve clinical diagnosis and Gato Weather prediction

treatment recommendation? Source Source
Source DeepMind.com PASQAL
Nam et al., 2022 A Generalist Agent The first neural model to
Nam et al. describe how multimodal AI systems could be implemented in Deepmind describes its Gato agent perform a diverse set of AI tasks
clinical routine as decision support tools. as “a multi-modal, multi-task, multi- These tasks include computer vision,
Focusing on histopathology and radiology data, the authors argue that by embodiment generalist policy.” image synthesis, vision-and-language
making use of various data types, AI could be advantageous in clinical settings. While most AI systems today qualify and NLP. Inputs will be unified and
as ‘narrow’, meaning they can converted to the chosen output. For
perform only a specific task, Gato instance, its image capabilities include
was developed as a ‘general- depth estimation, image inpainting, or
purpose’ AI that masters hundreds the detection of objects. In addition,
Clinical of skills – and might be a step Unified-IO users can give instructions
data
Genetic Medical towards Artificial General for newly generated images.
data history
Intelligence.
Diagnosis and
Pathology
slides treatment
recommendation

Radiology Temporal
images data
Laboratory
tests

93
Multimodal AI (2/2)
Assess
Data & AI

Key Insight
Google Searching with text and pictures

Multisearch In 2021, the tech giant Google introduced the

Multitask Unified Model (MUM). This model is now Multimodal AI is still immature but has great potential
Source implemented in the Google Multisearch via Lens. It has the capability to understand better the context of
Google This new product in combination with MUM conversations and therefore to support improving the customer
enables the search engine to process data in experience. Multimodal AI will enable us to combine text,
different modalities, like text or graphic, in a more speech, images and media into current context and derive
efficient and effective way than traditional appropriate conclusions and answers.
search methods.

94
Reinforcement Learning
Assess
Data & AI

Reinforcement learning Although significant progress has  RL can be used to estimate the
underwriting risk at minimal cost. It also
been made in the field, reinforcement
is an iterative learning learning is still mainly a research area. allows experiments to be run on live
quotation and sales data and can
process where a software therefore be applied to pricing strategy
agent faces a game-like optimisation

situation and employs

Risks
trial and error to find Opportunities  Applications for RL have previously been
a solution.  Reinforcement learning improves the limited by weak computer infrastructure
In reinforcement learning (RL), a quality of outputs from AI and data-  RL needs advanced computational
machine learning model is trained in based applications such as speech capabilities to be processed
an interactive environment by trial recognition for customer services
 Low sample efficiency: It takes the RL
and error, in which the model is  For insurers, claims management and system a considerable amount of time
rewarded for being right. Although the underwriting are major fields that to be efficient
designer sets the reward policy, i.e., the benefit from improved learning
rules of the game, the designer does not algorithms connecting real-world  RL is still a topic of research. We still
give the model hints or suggestions on events such as accidents to claims- await use cases with high relevance
how to solve the problem. Instead, the relevant data. In this way, manual for insurers
model uses feedback from its own processes involving a large amount
of previously unstructured data such
actions and experiences to improve. Its
as claims and contracts can
goal is to maximise the total reward.
be atomised to a higher degree
The models are designed to solve
 Like other deep learning technologies,
specific kinds of learning problems.
RL can be leveraged to prevent cyber
Fields of applications include self-
security attacks and threats associated
driving cars, industry automation, with persuasive computing
trading and finance, natural language
processing (NLP) or healthcare.
95
Reinforcement Learning
Assess
Data & AI

Typical framing of a Will RL NUON AI runs – and learns from –

reinforcement learning experiments on live quotation and
Environment
scenario
impact sales data

Similar to a robot finding its

insurance Insurtech startup NUON AI aims to optimise insurance
pricing by enabling insurers to adapt to market
way around a maze, each pricing? conditions in an instant – instead of waiting for pricing
and underwriting teams to respond to new market
Interpreter
random turn builds a picture entrants or competitors’ price changes.
of the layout until the robot Source
NUON’s Live Price Adjust AI product runs millions of
learns to navigate the maze. NUON AI experiments via API to make pricing strategies more
competitive.

Agent

RLHF
Bringing foundation models to Making driverless cars safer Key Insight
reinforcement learning and smarter
Source Source
DeepMind Phantasma Labs Reinforcement Learning Reinforcement learning
Conventional foundation models mainly Safe autonomous driving requires from Human Feedback learns through trial and
rely on self-supervised training. In testing at scale. To some extent, such uses human feedback as a measure
of performance.
error like human beings and
February 2023, Deepmind presented an testing can be executed virtually. The
doesn’t depend on historical
‘Adaptive Agent’ (AdA), a reinforcement- more diverse the data set, the better. Most recently, RHLF was used in
learning-agent that exhibits characteristics Phantasma Labs provides unique edge- data sets. First insurance use
ChatGPT to generate replies that
of a foundation model. AdA is able to case scenarios that involve complex sound more natural to people. cases are emerging,
adapt quickly to a large, open task space. vulnerable road user behaviour, saving especially in pricing.
clients the equivalent of hundreds of
hours on the road.
96
Artificial General Intelligence
Hold
Data & AI
 In theory, it could fully replace human
It may be possible to Expressed concisely and simply:
artificial general intelligence (AGI) – also
employees throughout the organisation

build a machine with called “strong AI” – does not currently  By combining advanced computing
capabilities with human-like self-reliant
exist, since today’s AI technology cannot
human cognitive be proven to possess the equivalent of
cognitive skills, AGI could help
companies solve the most complex and
human intelligence. It lacks common
capabilities – but we sense, intelligence and extensive
demanding tasks
 Take cyber risks: AGI might be the
are probably decades methods for self-maintenance or
reproduction. So today, AGI is merely a
appropriate technology to autonomously
prevent attacks, detect and fix security
away from developing subject for science fiction and “what if” leakages in software, and evaluate and
discussions. apply potential countermeasures, etc.
it. Progress on artificial intelligence has
so far been limited to so-called “weak Risks
AI”, in other words special purpose AI  Since AGIs don’t exhibit any human
limited to specific, narrower use cases. feelings like love or hate, they only aim
Even though it may be possible to build to achieve their goals, which, in turn,
a machine that approximates human may not be in line with human goals
cognitive capabilities, we are likely  Even for beneficial ends, an AGI might
decades away from having the be prepared to use malevolent means
necessary research and engineering.
 AGIs might be misused by unscrupulous
organisations, countries, etc.
 Depending on its development, AGI
Opportunities might be incomplete and misinterpret or
ignore contexts and commands
 Unlike other technologies, AGI doesn’t
require any technical know-how  Several open questions remain with
or highly-trained employees with respect to privacy, data security,
particular IT skills legal responsibilities, etc.

97
Artificial General Intelligence
Hold
Data & AI

The LaMDA controversy What are AI models When will unaided machines be able to
accomplish every task better and more
A Google engineer concluded that the AI ‘thinking’? cheaply than human workers? If we are to build AGI, we are
chatbot LaMDA was sentient. Experts soon Source
weighed in, declaring the assessment is false.
going to need to learn something
Anthropic from humans, how they reason
Founded in 2021 by Dario Amodei, and understand the physical
former VP of research at OpenAI, and 90% of experts gave a world, and how they represent
I want everyone to understand that his sister Daniela, Anthropic aims to date within the next and acquire language and
I am, in fact, a person. create “large-scale AI systems that are complex concepts.
100 years.
steerable, interpretable, and robust.”
Google’s chatbot LaMDA The company assesses current large, Gary Marcus
in a conversation that was released general systems as unpredictable, Half of the experts scientist, author, and entrepreneur
to the public in early June 2022 unreliable, and opaque. July 2022 in Scientific America
gave a date before
2061.
Key Insight
But AI researchers have been overly
optimistic before: Artificial general intelligence
doesn't exist yet and is likely
Of course, some in the broader AI community are considering the long-term decades away from being
possibility of sentient or general AI, but it doesn't make sense to do so by reached.
anthropomorphising today's conversational models, which are not sentient. Machines will be capable, within
Machines typically become
These systems imitate the types of exchanges found in millions of sentences twenty years, of doing any work a
and can riff on any fantastical topic. man can do. superior to humans only in
extremely specific tasks, but they
do not have the equivalence of
Brian Gabriel, a spokesperson for Google Herbert A. Simon
in a statement quoted in The Washington Post AI pioneer, in 1965 human intelligence.

98
Our Solutions
Data & AI
Munich Re Munich Re
aiSureTM aiSureTM and TWAICE
Guaranteeing the performance of artificial intelligence systems Battery analytics with a performance guarantee
aiSure™ is an insurance product for AI providers who wish to offer insurance- In 2022, Munich Re concluded the first AI green technology transaction for
backed performance guarantees. Available for AI start-ups or corporates, the Twaice's AI battery State of Health (SoH) estimation.
guarantees assure users that the AI service will perform as promised or provide TWAICE’s AI-based Battery Analytics Platform collects information relating
compensation if it does not. to the SoH of batteries. This is used to optimize the development and operation of
More than simply protecting users against underperformance, aiSure™ signals the batteries for electric fleets. The platform enables data-driven decisions to be made
provider’s confidence in the ability of its AI to perform. This increases the AI’s regarding asset management, operational costs and risks.
attractiveness to customers.

Benefits at a glance
 The AI provider benefits in its sales efforts as aiSure™ adds a competitive
advantage to its AI service
 The AI user gets a monetary indemnification that mitigates any potential Add a performance
system underperformance guarantee – and
 By insuring the performance of AI, Munich Re supports the adoption of customers can rely on
robust AI systems in businesses and society prediction-based decisions

99
Our Solutions
Data & AI
Munich Re Munich Re Munich Re
Insurance Analytics Platform Customer Analytics Building Insurance Analytics
From data to decisions Increase the value of each Data analytics for homeowner and commercial building insurance (ImRiSc)
Munich Re has created a platform for data customer and leverage sales What: ImRiSc identifies critical tap water risks for both commercial and private lines,
analytics solutions that enables their Customer analytics enables sales and and reveals the factors that drive risk in individual policies.
clients to fully exploit their portfolio data marketing professionals to generate How: ImRiSc combines internal and external data, machine learning and Munich Re’s
without investment into their own IT insights and gain a holistic view of each expertise, and provides appealing interactive visuals of the results.
infrastructure. The offering covers the insured based on analytical behavioral Where: ImRiSc supports (1) a client’s portfolio restructuring and (2) the evaluation of
whole primary insurance value chain. models. new business.
Two examples of solutions are Customer Benefits at a glance  New business: ImRiSc evaluates risks solely on the basis of external data with only
Analytics and Building Insurance  Customisation of solution to the the risk address required
Analytics (ImRiSc). clients’ lines of business and system  Transparency: ImRiSc comes with an interactive dashboard to identify risk drivers
Benefits at a glance landscape and explain the AI-driven results
 Strong commitment to improve your  Combining a retention, a cross-selling Benefits at a glance
performance and an overall customer experience  Superior statistical benefits: massive uplift in predicting future tap water claims
 State-of-the-art analytics services with module  Huge data base
highest IT security and Trustworthy AI  Holistic picture of the entire customer  Internal data consist of the client’s proprietary data and anonymized data from
standards for GDPR-relevant data relationship through Customer Value a variety of portfolios
 Service offerings in various areas like Score  External data supplements and enriches internal data
Underwriting Monitoring, Claims,  Interaction of a data-based complete
Customer Analytics, Big Data and picture of the customers and a whole
Artificial Intelligence – see for example range of predictive models
Customer Analytics and ImRiSc offering  Constant further development based
 Easy web-based access to your on new methodologies, findings and
customized analytics results, with clients’ needs
possibility to consume results as APIs  Illustrative and interactive dashboard
to identify high-value customers

100
Our Solutions
Data & AI
Munich Re  Strengthen their competitive position Munich Re
IMPROVEX  Open up new possibilities to identify P&C insurance pricing and UW
Data-driven excellence for your attractive business potentials Scaling prediction-building with AutoML
portfolio Global Consulting offers support across the whole of end-to-end insurance pricing and
IMPROVEX is Munich Re’s innovative underwriting for clients. Scalability is enhanced by Akur8 – a SaaS modeling workbench
data-pooling platform that adopts a for risk, behavioral and market pricing. Munich Re has an umbrella contract that allows
unique approach by smartly combining us to boost our internal delivery capabilities.
claims, premium and exposure information  It improves insurers’ processes by automating risk and demand modeling using
with Munich Re’s renowned actuarial artificial intelligence
expertise and advanced data analytics.
 It dramatically reduces the time required for modeling
By aggregating and analysing our own
policies and claims data and those of We tailor our products to  It provides an out-of-the-box, risk-smoothing approach for motor and
home portfolios
selected participating clients in a our partners’ needs by
comprehensive database, we create The prediction-building process is traditionally iterative. Akur8 delivers better results in
sharing market insights. a fraction of the time by leveraging AutoML and allows actuarial-driven decisions on
unrivalled market insights and boost our
impact/dislocation analyses, profitability and price optimization.
clients' profitability in a challenging One of the most important
industrial risk environment.
success factors of
IMPROVEX delivers client value via a
two-pillar service model
IMPROVEX is the close
 Web App for delivery of core co-creation with our
actuarial products clients.
 Actuarial consultancy and
further analyses
Katja Steinmann
With our solution our participants can ...
 Permanently improve their underwriting Head of IMPROVEX
and growth strategy
101
Our Solutions
Data & AI
Munich Re Munich Re
RE_Chart RE_Wheel
Pricing enhancement based Pricing sophistication via vehicle allocation
on POIs We deliver your Vehicle features offer one possibility to enrich data and improve risk assessment.
Homeowners’ portfolios in mature competitive edge by Yet the large number of features leads to high complexity of the modeling process.
markets are typically profitable, but leveraging your data, RE_Wheel offers state-of-the-art machine learning algorithms and proven methodology
margins in building covers are thin. Our to support the addition of vehicle details to models.
cedants can tap into non-standard ways technology and
Two levels of sophistication are possible
of boosting the technical sophistication of our expertise.
 Vehicle mapping: spatial representation, with similar vehicles in neighboring regions
risk modeling.
Personal lines portfolios, from  Vehicle allocation: various distances between vehicles derived, vehicle clustering
Risk Network Analytics (RNA) is a and improved modeling
technique that allows the remoteness of
motor to property to travel, are
any risk from a set of chosen points of traditionally the most analytic-
interest (POIs) to be taken into account. driven ones and most prone to
Munich Re
Munich Re leverages umbrella contracts hyperautomation.
for this data enrichment, but applies RE_Map
proprietary IP for the insights This is where Munich Re brings
Microzoning for P&C insurance
a unique combination of global
 RE_Chart is a scalable solution that Geo-related risks benefit from the availability of large datasets. In this context, machine
processes the shortest distance exposure, subject matter expertise
learning techniques and the use of external data are continuously gaining in relevance
between a portfolio risk’s location and and advanced analytics solutions
and can foster competitive advantages. Leveraging advanced machine learning
each point of interest combined with risk transfer. techniques and Munich Re’s unique data assets, RE_Map allows the geographical
 When a bespoke solution is needed, the component of the cost of the risk to be assessed optimally.
full power of RNA is used (complex
routing techniques, real road network) The general approach involves 3 main steps
Dr. Davide Burlon
 External data enrichment in risk and behavioral modeling
Principal  Residuals smoothing correction
Insurance Solutions  Classification into zones with homogeneous risk

102
Our Solutions
Data & AI

Growing digital
motor insurance in
Southeast Asia:
Applying customer-
centric and data-driven
propositions in one of the
world’s most vibrant
regions.

103
Our Solutions
Data & AI
Munich Re Munich Re
FIVE Talaria
Rule-based investment strategies Empowering digital trade finance
Using modern data-science techniques, Talaria is a fast-growing trade finance
FIVE develops rule-based investment venture from Munich Re that works with
indices as performance engines for partners to integrate white-labelled
savings and retirement products. invoice-to-cash solutions for SMEs into
Insurance clients value the convenience their platforms. With the objective to
of a one-stop shop that combines increase customer satisfaction and business
investment strategies, guarantees and potential, Talaria offers your client base of
insurance covers into a single product small and medium-sized companies a
solution – lean and tailored to individual simple way to optimize their working
needs. Moreover, they can access a suite capital and enhance their cash flow.
of highly diversified return sources, With Talaria, you can co-create your
benefit from cost efficiencies, and aim to specific solution suite to enable your
improve their balance sheet utilization. customers to take out credit insurance
Accessible via reinsurance with and monetize their invoices with just a
Munich Re Markets. few clicks.
Benefits at a glance Benefits at a glance
 Access to a selection of quantitative  Plug and play white-label solution
investment strategies  Immediate risk assessment and
 Better risk transfer by sourcing cover decision
complete investment solutions  Fully compatible API
directly from Munich Re  100% GDPR-compliant
 Attractive payouts of guarantees  AI-based superior scoring model using
and insurance covers actual trading data
 Automatic management of limits

104
Cyber & Crypto
Technology impacts in 2023

European
Commission
Reduction of
aims to have
99.95%
in total energy use
per transaction with 20%
80%
Ethereum’s PoW. of EU citizens using a
of account takeover attacks
in 2023 will be made digital identity system
53% with deepfakes
help.
The global healthcare
by 2030.
of cyber security
organisations prioritise 60% biometrics markets is
predicted to grow to
“ensuring better of all corporate data

US$ 32.9m
privacy of is stored in
the cloud.
customer
data”. by 2030.
All trends 2023
Cyber & Crypto
The trend field of Cyber & Crypto is home for technologies around secure transactions and secure For insurers, new technologies in the realms of privacy-enhancing technologies and automated
ways of transferring data. Data encryption is a key technological principle, and blockchain technology compliance change the way data is transferred and handled between clients, client managers, and
is the enabler for trends in this field, such as digital assets and smart contracts. Leveraging these underwriters, for example. AI-manipulated pictures, videos or voices, so called deepfakes, continue to
technologies, decentralised data economies and ecosystems are being built. Building trustworthy be a conspicuous cyber threat. Deepfake defense solutions, along with human recognition
digital solutions with enhanced data protection is crucial in cyber security. technologies that allow for secure authentication, are further evolving.

Adopt – Start initiatives in your unit Trial – Initiatives in affected units Assess – Evaluation needed Hold – Watch list
Cyber Security Deepfake Defense Automated Compliance
State-of-the-art prevention mechanisms and quick Technological solutions to combat the spread of As the significance and scale of data increases
incident responses help companies to be one step ahead manipulated information. Semantic techniques help incessantly, awareness of privacy and other compliance
of cyber-attacks like insider threat, DDoS attack and detect and remove highly sophisticated and risks has taken on a central role for businesses.
E-mail fraud. believable AI-manipulated media known as deepfakes. Especially in the insurance sector navigating the world
of compliance is complex and resource-intensive.
Digital Identity Digital Currencies
Identification represents a necessary basis for Digital currencies are a group of digital assets and refer Decentralised Data Economy
transactions, not only for individuals, but also for to currencies that are generated, stored, and transferred The decentralised data economy is the cradle of an open,
organisations, governments, and objects. A digital only in digital or electronic form. Triggered by blockchain transparent, decentralised digital ecosystem where data
identity enables digital transactions and today is technology, decentralised cryptocurrencies, virtual and services can be made available, collated, shared and
tied to a plethora of daily-life applications and data. currencies and central bank digital currencies are monetised in an environment of trust.
gaining traction.

Human Recognition Privacy-enhancing Tech

Modern human recognition systems utilise intrinsic Privacy-enhancing technologies (PETs) represent a wide
(physiological and biological) and extrinsic (behavioural) range of hardware or software solutions designed to
human characteristics, allowing highly secure maximise the social, scientific and commercial value of
identification and authentication with real-world data, incorporating data protection principles through
applications. minimum data use, higher security and enhanced privacy.

Smart Contracts
Based on programming language smart contracts
enforce agreements via software rather than law.
Therefore, they use digital tokens which represent all new in 2023
kind of real-world and digital assets.

107
Trend Evolution
Cyber & Crypto

Evolution over time

Adopt Trial Assess Hold

Start initiatives in your unit Initiatives in affected units Evaluation needed Watch list

Automated Compliance 2023 2022

Cyber Security 2023 2022 2021 2020 2019

Decentralised Data Economy 2023 2022

Deepfake Defense 2023 2022 2021 2020

Digital Currencies 2023 2022

Digital Identity 2023 2022 2021 2020 2019

Human Recognition 2023 2022 2021 2020 2019

Privacy-enhancing Tech 2023 2022

Smart Contracts 2023 2022

108
Impacts on the Value Chain
Cyber & Crypto

Explanation Insurance Value Chain

The illustration represents the impact the trend
fields have upon on the insurance value chain. Product Sales & Risk Customer Asset
With the dark boxes reflecting areas of high Design & Pricing Distribution Underwriting Policy Admin Management Engagement Claims Management
impact, and light boxes low impact.

Automated Compliance

Cyber Security

Decentralised Data Economy

Deepfake Defense

Digital Currencies

Digital Identity

Human Recognition

Privacy-enhancing Tech

Smart Contracts

109
Cyber Security
Adopt
Cyber & Crypto

Leveraging the opportunities of digital connectedness unfortunately comes with Risks

increased vulnerabilities for individuals and companies. State-of-the-art prevention  As cyber security solutions mature,
mechanisms and quick incident response can help companies be one step ahead they will become vulnerable to
next-generation hackers
of cyberattacks.
 Reliance on cyber intelligence may
IT systems are becoming ever more complex, pervasive, and crucial for give institutions a false sense of
our way of living and working. The evolving Internet of Things, remote Opportunities security against cyberattacks
working, as well as the proliferation of cloud computing, broaden the  New technologies like cognitive
 Cyber security systems that use
scope of vulnerabilities. Cyber criminals are poised to exploit these capabilities featured in security
cognitive technologies enhance
vulnerabilities – at a high cost for their victims. According to a report software, are not yet established as
the early-warning systems of
by Cyber Security Ventures, cyber crime is going to cost the world reliable and trustworthy cyber
IT security departments
US$ 8 trillion in 2023. Measured as a country, cyber crime would be security solutions. Highly
 Smart cyber security enriches specialised resources that combine
the world’s third largest economy after the US and China. and strengthens the service the fields of cognitive technologies
Given this outlook, cyber security solutions that support the portfolio of organisational and and cyber security first need to be
prevention, detection, and examination of attacks must advance fast. industry-related cyber security built and convincing investment
From cognitive to AI-powered, the range of technologies ready to be  Smart cyber security systems plans drawn up. Operational costs
leveraged for cyber security is broad. may enable insurers to better may increase substantially
understand and control cyber
For instance, novel cyber security systems integrate cognitive risks and therefore underwrite
capabilities to “understand” information from external sources and higher limits of cyber coverage
internal systems to identify cyber security risks. Machine learning
mechanisms not only help detect attacks; they also speed up
the examination process to make the right decisions after notification.
110
Cyber Security (1/2)
Adopt
Cyber & Crypto

Which cyber security Top five priorities of organisations in cyber security

Martin Kreuzer
threats do CISOs expect? worldwide in 2022
Senior Risk Manager
Cyber Risks
Better ensuring the privacy of customer data 53%
1. Insider threat
Implementing or enhanding a zero trust model 51%

2. DDoS attack Better addressing partner and supplier risks 43%

From a Cyber Underwriting
3. E-mail fraud Better measuring effectiveness of our security program 36%
and Risk Assessment
Achieving industry or regulatory compliance 34% perspective it is of utmost
4. Cloud account compromise importance to closely monitor
cyber security trends and
5. Malware developments. It is key to
HUB Security
know about growing digital
6. Ransomware attacks Source dependencies and how these
7. Smishing/Vishing hubsecurity.com are exposed to cyber threats.

8. Supply chain attacks Confidential computing At the same time, we need

Data security has traditionally focused on protecting data at rest and in to know and thoroughly
transit. However, during processing, data often remains vulnerable as it is assess how policy holder
difficult to protect. Tel Aviv-based company HUB Security allows data to and applicants may protect
remain encrypted even as it’s being processed. themselves in a risk adequate
way!

111
Cyber Security (2/2)
Adopt
Cyber & Crypto

Munich Re Global Cyber Risk and Insurance Will ChatGPT impact

Survey 2022 cyber security?
In many ways, artificial intelligence
 83% of all C-Level respondents globally report that their company is not (AI) technologies have contributed
adequately protected against cyber threats to improved protection against
cyber attacks. For instance,
 The number of C-level participants who are „extremely concerned“ has
machine learning and natural
increased from 30 to 38%
language processing are commonly
integrated into cyber security
Why does your company have no cyber insurance solutions. Yet, evolving AI
applications like ChatGPT may be
in place?
misused to the benefit of the cyber
criminal, be it through enhanced
phishing mails or the creation of
Price for coverage was too high 29% polymorphic malware.

Did not know that cyber

insurance existed
25%

Did not understand the product 22%

Perceived the scope of the service/

18%
coverage as insufficient
Cyber & Crypto
Cyber Security Highlight
Emerging Cyber Security Trends
Adopt
Cyber & Crypto

Physical Data Security Application Creator Economy

Cyber Security Posture Management Security Cyber Insurance
 In recent years, software-related  The adoption of cloud infrastructures  Digital transformation initiatives have  There are approximately 50–100m
security products have superseded has transformed the way transformed the enterprise application social media influencers globally
physical (hardware-related) cyber organizations interact with data, security landscape. To facilitate that monetise their content. These
security products in terms of preferred which leads to moving, duplicating or virtually all digital interactions within individuals are part of the creator
cyber security solutions for security changing sensitive data assets. Most an organisation, the proliferation of economy, a label that is inclusive of
teams. However, uncontrolled hardware organisations and CISOs do not know influencers, gamers, bloggers and
internal and external applications and
assets still lead to data loss, system where their sensitive data sits and community builders
APIs is essential
downtime, business disruption and – therefore cannot properly protect  What was previously missing, but is
for some organizations – major their most valued assets  APIs serve as the intermediary that now starting to emerge, are financial
reputational damage  Companies are increasingly dealing enables applications to speak to one management tools and services built
 Physical cyber security allows teams with decentralised data stores and another; a common analogy is that specifically for the creator economy
to evaluate and isolate critical unstructured data. This data chaos APIs are the doors and windows  Creators are professionalizing their
infrastructure, information, and creates shadow data vulnerabilities (that need to be locked) of an operations by taking more ownership
data stored on devices to prevent and governance breaches leaving enterprise security perimeter. of their digital assets and follower
malicious actors from entering the companies exposed because of experiences
company’s environment planned cyber-attacks or human  Enterprises face myriad challenges
error. On average, the two biggest related to application security: inability  Today digital SMBs and creators are
factors contributing to the cost to observe machine-to-machine underserved and have no low-cost,
of a data breach is a complex interactions, insufficient engineering zero touch insurance options tailored
security system and cloud migration resources, a lack of visibility into specifically to protect their accounts

 Data Security Posture Management application dependencies and scaling  As many more decide to pursue
(DSPM) solutions better understand challenges to meet the demands of a creator career and more small
where data sits, if and where it is modern organizations businesses use social media for
exposed, who has access and, most marketing, their livelihoods
 Modern application security should become exposed to hacks
critically, how to improve and
address the culmination of these risks
remediate security posture
to secure enterprises

114
Asset Risk Management
Adopt
Cyber & Crypto

Asset Risk Management becomes a growing

challenge in an interconnected world
Yossi In the rapidly expanding, uncontrolled ecosystem of hardware, companies have a
difficult time to track their assets and identify potential risks. Often, IoT hardware
Appleboum can be more accessible to hackers or insider threats than standard PCs:

By 2025, it is expected that there will be more than 30 billion IoT connections,
Co-founder and CEO, Sepio almost 4 IoT devices per person on average and that also amounts to trillions
of sensors connecting and interacting on these devices. (iot-analytics.com)

127 new devices connect to the internet every second. (mckinsey.com)

So many companies are focused on cyber security, but we

specifically tapped into asset risk management from the beginning Bolstering Cyber security in the Physical Realm
when this was an inexplicably overlooked area. It has truly paid off.
So many great minds are now recognising asset risk management With Sepio’s Asset Risk Management (ARM) platform, customers
as a key cyber security challenge. have a complete “zero trust” architecture of their infrastructure
Sepio sees, assesses and mitigates all known and shadow assets, regardless of their
that uncovers all connected assets, verifies their true identity,
interface (USB or network) or functionality (IT/OT/IoT), as fast as they are added by and provides asset risk management.
anyone, anywhere. Focused on device existence rather than activity, our unique
approach offers holistic asset visibility based on truth, identifying every asset’s
“DNA” using our patented algorithms. Untainted by misleading profile perceptions, Sidra Ahmed
behavioral assumptions or inactivity, Sepio gets to the true source of asset risk with
minimal resource requirements, providing customers with unparalleled asset risk Investment Principal,
management that is infinitely scalable. Munich Re Ventures
115
Data Security Posture Management
Adopt
Cyber & Crypto

Strengthen your cloud  Cloud security professionals estimate

that 50% of their sensitive cloud data
data security posture is not properly secured. In reality most
 Discover, classify and track your don’t know where their cloud data is
Asaf sensitive data including source code and developer
secrets
Kochan  Understand who has access to
 The average cost of a data breach is
sensitive information
expected to reach $5m in 2023
Co-Founder and President, Sentra  Detect data duplication and movement
 Defensive security controls are
 Notify data owners when a sensitive insufficient to protect movement
asset’s security posture changes of sensitive data in the cloud
 Provide context and actionable steps
for quick remediation

As data is generated in the cloud, security and compliance

requirements multiply. This is hurting business processes the Providing DSPM in the cloud
cloud was designed to improve. By focusing on valuable data,
enterprises can better deploy their resources to keep the business Increased visibility and control over shadow data will improve internal
secure while taking advantage of all the cloud offers. data utilization, reduce data management costs and address growing
systemic risk. We believe Sentra’s unique insight into the shadow data
Sentra has pioneered Data Security Posture Management (DSPM) - a new, data-
centric approach to natively secure all of your data as it moves through your cloud landscape can ultimately inform data governance risk assessment for all
environment. We instantly bring all of your sensitive cloud data to light, classify it by industries, including cyber insurance.
sensitivity, and pinpoint the specific steps to remediate data vulnerabilities fast.
Due to the growing market need to detect, classify and secure moving Veronica Gabriele
data in the cloud – Gartner first recognized Data Security Posture
Why now? Senior Investment Associate,
Management (DSPM) as an emerging trend in cyber security in August
2022. (gartner.com) Munich Re Ventures
116
Modern Application Security
Adopt
Cyber & Crypto

By 2026, 40% of organizations will select their 321%

web application and API protection provider
YoY increase
based on advanced API protections, and web
in overall
Greg application security features — up from less
API traffic
than 15% this year.
Martin
Co-Founder and CEO, Ghost Security
By 2025, less than 50% of enterprise APIs will be 681%
managed, as explosive growth in APIs surpasses increase in
the capabilities of API management tools. API attacks

The explosive growth of microservices and APIs in the

cloud will continue to outpace the capabilities of An innovative approach to Modern
existing security solutions. Application Security
Ghost is establishing the new standard in application security for
We see this creating an even larger gap between the defensive capabilities
enterprises undergoing (or completed) digital transformation initiatives.
of enterprise businesses and the capabilities of skilled hackers.
Ghost was founded to stop application breaches by helping organizations
At Ghost, we are completely rethinking the approach to securing modern discover, detect and defend against modern AppSec threats.
applications from the ground up. The team at Ghost is uniquely qualified
to develop new innovations in cloud security, specifically securing
Blake Pennington
vulnerable applications, microservices, and APIs.
Senior Investment Associate,
Munich Re Ventures
117
Insurance for Digital Assets
Adopt
Cyber & Crypto

There are over 31.5m businesses with under 20 employees in

the U.S. More than 85% of those micro businesses don’t buy
cyber insurance, even though they face the same risks.
Rafael
Broshi
Co-Founder and CEO, Notch
7 out of 10 businesses in the U.S. use social media as part of
their business, for brand awareness, distribution and
customer communications.

Social engineering attacks will be the number one

cause of cyber attacks over the next few years, meaning Providing much needed cyber coverage
that a new approach must be taken to assess risks, for SMBs
even at an organization level. Notch is seeking to create a seamless underwriting process with
automatic recommended policy generation. The company is developing
We believe that cyber threats should be managed from an individual a simple and seamless business continuity offering that gives creators
perspective. Over the last 6 months, we have launched the world’s first and their sponsoring partner brands peace of mind that their
coverage for social media accounts against account takeovers. assets and income are protected.
By starting bottom up and insuring specific digital assets, we are able to Oshri Kaplan
uniquely identify risks individuals face.
Investment Managing Director,
Munich Re Ventures
118
Digital Identity
Adopt
Cyber & Crypto

Identification represents a Opportunities Risks

necessary basis for transactions,
 Digital identities reduce fraud risks,  Although concepts of privacy differ
not only for individuals, but also since they are much more difficult across cultures, countries require
for organisations, governments, to copy or fake than paper-based a stable framework for data protection
and objects. A digital identity documents that covers the storage, linkage and
use of data
enables digital transactions and  Consumers can save time and effort,
today is tied to a plethora of daily- since customer onboarding and  Digital identities need to be protected
identity checks (particularly in from theft and abuse
life applications and data. the financial industry) only need to be  Digital authentication requires high
To verify the authenticity of a subject in a done once privacy protection to mitigate risks
digital setting, a single or a combination of  Consumers may control their own of unauthorised access to individuals’
identifiers is used. These identifiers can be digital identity and decide which information
static – such as an e-mail address, a elements of it are disclosed to which  Digital identities may be faked in
password or a serial number on a product – counterparties. Health insurers, for order to pretend to be someone else
or dynamic, including entity relationships instance, only need access to or may be hijacked as a type of
captured from the user’s behavioural customers’ health information, but not security attack to control identity
history. to their overall financial situation
The dynamic methods of digital  Risk assessment for insurance covers
identification, in particular, promise higher may become entirely based on digital
accuracy and tend to be less prone to fraud identities. In health insurance, for
and theft. Still, the topic of digital identity example, a patient’s complete health
raises concerns regarding safety, privacy record would make extensive
and certainty. Technology companies along questionnaires obsolete
with public actors seek to optimise digital
identity management by creating
streamlined, secure and trustworthy
processes in addition to user-centric and
user-controlled data frameworks.
Digital Identity
Adopt
Cyber & Crypto

BYOI
 It is inefficient and costly to have to prove one’s
identity over and over again
 In the BYOI model, a user’s identity data is stored
Bring Your Own Identity –
locally in the user’s personal digital wallet
merging identification
 Gartner predicts a global standard for decentralised
and authentication
identities by 2025

In Hong Kong, digital life insurer Blue AIA Singapore connects to the
has adopted the government’s iAM government’s Singpass program, The European Commission aims

80%
Smart authentication method. which enables users to access more to have 80% of EU citizens
Registered users can verify their than 1,400 government and private- using the European digital
identity via the iAM Smart mobile app. sector services online and in person.
identity system by 2030.

Two categories make up one’s digital identity Digital ID wallets: safer and
Key Insight
smarter in the future
Digital attributes  The Covid-19 pandemic has pushed
such as bank details, the adoption of digital ID wallets
Digital identities can render
social security number, insurance processes more
 A new generation of digital ID wallets
medical history, efficient and secure.
is adding scanning and document
government-issued ID. Digital activities
encryption to its capabilities Companies should look out for
such as purchase history,
 Future digital ID wallets will store new use cases as increasingly
downloaded apps, mobile
phone usage, search essential documents, payment more confidential data is
queries, social media likes. information, medical history as shared, e.g., in health records
well as biometric information or claims management.

120
Human Recognition
Adopt
Cyber & Crypto

Modern human recognition Opportunities

systems utilise intrinsic  Fraudulent insurance claims cost billions of dollars annually. The human recognition
(physiological & biological) system is an effective way to combat fraud and reduce the ensuing overall costs
and extrinsic (behavioural)  This transformative technology can be used for processing and assessing insurance
human characteristics, claims to determine their validity and reduce processing time
allowing for highly secure  Facial recognition, together with other personalised information, allows for more precise
identification and risk detection and tailored quotes that are more reflective of the actual levels of risk
authentication with  The use of machine learning and human recognition systems may lead to the
identification of previously unknown risks, resulting in better risk prediction, refined
real-world applications.
risk management and lower cost
 The shift to AI and the use of human recognition systems can lead to credible
Biometric-based recognition systems
underwriting, time-saving and easier claim settlement
overcome the security vulnerabilities of
conventional knowledge-based
methods. They thus enable applications
to meet strict security requirements. Risks
Amazon Recognition and Microsoft’s
 Data confidentiality is a primary concern in relation to human recognition systems.
Computer Vision API are among the
These relate to measures to prevent data breaches, unauthorised use and access, as
human recognition systems that offer
well as the preservation of data integrity
identification capabilities to extract
information from images and videos,  Privacy violation, claims of gender and racial bias and lack of industry-wide
with applications in physical security, regulations in the use of AI are also among the other main hurdles to the use of
cyber security, military, employment, recognition systems
insurance, financial transactions, public  This is an emerging technology still in need of greater development and oversight for
services and criminal justice. However, deployment in production. Low accuracy in claim processing and validation arose in
these services have come under previously deployed recognition systems. Furthermore, claims of ableism and racial
continuous scrutiny, primarily due to a bias suggest that the deployment of poorly-developed recognition systems
lack of regulations. may encourage unintended discrimination

121
Human Recognition
Adopt
Cyber & Crypto

Extrinsic biometric Intrinsic biometric Digital Overcoming privacy risks in the storage of
facial images in medical records
characteristics characteristics
include gait, voice, keystroke include fingerprint, face, hand,
Masks Clinical data involving biometric identifying information,
especially facial images, are especially sensitive. In order
dynamics, mobile interactions, iris, ear, vascular patterns, Industry to reduce privacy risks, researchers are looking for ways
signature and mouse dynamics. electro-cardiogram and DNA.
Healthcare to erase identifiable features while retaining disease-
relevant features needed for diagnosis.
Source Based on three-dimensional reconstruction and deep-
Yang et al. 2022 in learning algorithms, a team of researchers in China has
Nature Medicine developed a so-called digital mask, which achieves very
high diagnostic consistency and accuracy.

Face and voice recognition technology to 2030

Key Insight
transform approach to perinatal depression CAGR
Source
BlueSkeye AI (University of Nottingham) +24% 2022 Human recognition
enhances opportunities in
University spin off company BlueSkeye AI has developed clinical
fraud and risk detection.
grade face and voice analysis technology which uses machine
learning to help assess health, mood and mental state. Its first Companies should at least
Estimated growth of the global
application is an app, Avocado, which uses the technology to help initiate pilot projects and
women from 3 months pregnant to six months post partum
healthcare biometrics market,
evaluate the value chain for
monitor and manage their mental health. predicted to surpass
use cases.
US$ 32,975m by 2030.

122
Deepfake Defense
Trial
Cyber & Crypto
 Better fraud detection: The ability to
Technological such as clinical research, problematic
detect distorted or AI-generated photos
applications are evident. Risks include and videos reduces the risk of digital
solutions to combat the market manipulation, cyberbullying, media fraud in insurance claims
spread of manipulated blackmail and political instability.
 Emphasis on veracity safeguards can
Moreover, misuse of deepfakes may create a company culture where
information, such as AI- erode trust in media in general. people are more aware of threats and
manipulated media known With the viral spread of deepfakes, the less likely to fall prey to false information
need for scalable systems to recognise
as deepfakes. and prevent fake information is rising. Risks
In a recent report, researchers from Deepfake detection techniques, such as
 Statistical methods to validate a fact are
University College London ranked AI linguistic analyses or spatial and
overruled by deepfake manipulation
applications that could be used to temporal clues have become a hot topic techniques. Even (untrained) human
facilitate crime over the next 15 years. for researchers and technology experts. review is not reliable when it comes
At the very top, they placed deepfakes. Beyond detection techniques, to detecting deepfakes
Deepfakes refer to media – texts, approaches of deepfake defense include  Deepfakes have made their way from
pictures, videos, audio or data sets – that identity watermarking or the humorous fake news or a potential
were synthetically created or digitally attachment of identifying labels. Yet, national security concern to faked
altered but appear to be ‘real’. While technology experts and lawmakers find corporate or client communication.
deepfakes and synthetic data hold great themselves in a constant struggle to Dedicated and regular training not only
promise in some areas, outpace deepfake creators. for IT security staff but also for client,
key account or demand managers
is essential
 Training material and professional
Opportunities deep-fake detection is still not
widely available
 Increased accuracy in claims
management using recognition  The scalability of semantic deepfake
systems detection techniques for data-rich
systems such as social media are still
in their infancy
123
Deepfake Defense
Trial
Cyber & Crypto

Content Authenticity Initiative (CAI) TruePic Factmata

Source Source Source
contentauthenticity.org truepic.com factmata.com
Gartner predicts Authenticate digital The future of social analytics
Digital content provenance that 20% of
Founded by Adobe, Twitter and the New York Times in successful content at scale Factmata has developed algorithms that
account take- Truepic helps companies collect trusted, are capable of classifying content and
November 2019, the CAI is creating a secure end-to-end system
over attacks in authenticated content through its fraud detecting fake news. Via linguistic cues, the
for digital content provenance through open-source development,
2023 will be prevention mobile camera integration technology spots misinformation that was
cross-industry collaboration, and interoperability of tools.
made with the technology. written to sound authentic and truthful.
help of deep-
fakes.

What is a ‘cheapfake’? Some ‘cheapfake’ indicators

The term ‘cheapfakes’ refers to
In pictures, deepfakes have
audiovisual manipulations of videos
and photos created with simple means.
irregular pupils The best defense against deepfakes is a
In contrast to technologically far more Deepfake audio has subtle vigilant and informed society that
differences in the high
sophisticated deepfakes created by
frequencies recognises the potential for malicious use
experts with the help of tailored artificial
intelligence and machine learning, and takes steps to mitigate its impact.
In videos, deepfakes
cheapfakes can be produced using easily
do not blink
accessible methods and conventional Dr. Stephan Meyer
programs. But even for cheapfakes,
detection is like a 'cat and mouse game’ Text deepfakes use the
Head of AI
most common words
as apps become more advanced.
124
Digital Currencies
Trial
Cyber & Crypto

Digital currencies are a group One subset is digital currencies issued  Some crypto coverage start-ups price volatility but even they can be
of digital assets and refer to by a community network. Bitcoin gather cryptocurrency from crypto subject to collapses
currencies that are generated, represents a decentralised, blockchain- investors to use it as insurance capital.  Since no intermediary is involved in
based digital currency and has Cryptocurrency can either be used in many digital currencies, individual
stored, and transferred only in digital
exceeded market cap of US$ 1tn. the same way as traditional insurance parties now face risks that were
or electronic form. Triggered by Similarly, the cryptocurrency Ether capital or can be employed as a stake previously managed by a central
blockchain technology, decentralised has reached heights of over in the pricing of risk authority
cryptocurrencies, virtual US$ 250bn.
 The emergence of digital currencies
currencies and central bank On the other hand, central banks are Risks gives rise to legal, regulatory, and privacy
digital currencies are gaining traction. looking into digital currencies and will
 Digital asset owners have crypto issues. International agreement on how
While the idea of currency in digital form likely soon launch their own digital to handle these issues is lacking to date,
currencies, combining the benefits of wallets containing the private keys
goes back some 40 years, more recent even if regulation is evolving steadily
cryptocurrencies and traditional money. needed to access their crypto assets –
efforts to apply emerging technology in
assets, which exist on the blockchain.
currency creation represent a remarkable
Therefore, the private keys tend to be
innovation in payment systems. Beyond
susceptible to hacks: Although private
known concepts of mobile money or e- Opportunities wallets are generally very safe,
money, novel ledger technologies have
 Digital currencies can facilitate hackers only need to steal a user’s
helped overcome previous obstacles to
international transactions, offering private key to tie the ownership of
digital cash.
high speed and low costs digital assets to a different wallet. This
Today, digital currencies vary by issuer means that the possession of private
(such as central banks, private  Stablecoins are designed to reduce
keys also confers ownership of
organisations, or a community network), volatility. They are pegged to a
coins, data, content, etc.
value backing (claim-based or object- cryptocurrency, fiat money, or to
exchange-traded commodities  Some digital currencies exhibit high
based), supply mechanism (governance
volatility. These high value fluctuations
decisions on whether and how to adjust  As more value is held in the form of
challenge their capacity as a medium
supply), and their ledger technology (which digital currencies, demand for
of exchange. Asset- or fiat-pegged
can be publicly accessible or private). insurance solutions rises. Crypto wallet
stablecoins are designed to minimise
coverage is an opportunity for insurers
Digital Currencies
Trial
Cyber & Crypto

European Commission’s proposed With evolving regulation and more

institutional asset holders entering the
Markets in Crypto-assets (MiCA)
crypto economy, a strong demand for
directive will be a substantial piece regulated insurance capacity for crypto
of regulation for crypto-asset custody solutions is facing limited supply Johannes
service providers (CASPs) in the by specialised insurance carrier. With
its Digital Asset Protection Custody Benedikter
EU. It aims to harmonise the
Cover, Munich Re can provide suitable
European framework for the Senior Underwriter (Digital Asset Protection)
insurance backing that will support the
issuance and trading of various safeguarding promise given towards New Tech Underwriting
types of crypto token as part of these institutional clients. The cover
protects against theft and permanent
Europe’s Digital Finance Strategy.
loss of digital assets, caused by
The directive has four general and for example
related objectives
 To create legal certainty, clearly defining Within our dedicated Digital Asset Protection team, we
the regulatory treatment of all crypto are providing several risk transfer solutions for the crypto
assets that are not covered by existing
financial service legislation, economy. Our aim is to bring security in this yet emerging
 To promote the development of crypto- external physical insider new market.
assets and the wider use of distributed hacking theft crime
ledger technology,
 To instill appropriate levels of consumer
and investor protection, and
 To ensure financial stability
Amongst other things, this directive defines
Q2-23 Q1-24 Q4-24 25/26
MiCA expected Provisions for Asset MiCA applies EC report on
prudential requirements for CASPs, which
to become law Referenced Tokens & (full application 18 months MiCA application
also takes risk transfer solutions into account. E-money Tokens apply after becoming law)
126
Smart Contracts
Trial
Cyber & Crypto

We are all familiar with contracts which you can copy and paste without caused by events in the DeFi  In addition to cyberattacks, third-party
loss of quality). Tokens can either be industry increases demand for exploitation of blockchain protocols and
written in natural language. Smart
interchangeable (fungible in that one insurance products partial lack of security standards are a
contracts, on the other hand, are unit does not differ from another) or concern. Some of the largest hacks
written in programming language  Tokenisation yields opportunities
unique (non-fungible tokens or NFTs). were carried out in the DeFi space, e.g.,
regarding flexibility (e.g., in terms of
and enforce agreements via fragmentation) and transfer of
on bridging services, which connect
different blockchain networks to
software rather than law. They are asset ownership
facilitate cross-chain transactions
also the basis of tokens that can
Opportunities  The transparent nature of smart
 Lack of standardisation and differences
represent all kinds of real-world contracts’ code is creating a new field
in methodology and coding techniques
 Automatic execution of agreements for auditing companies. Quantstamp,
and digital assets. via smart contracts offers various hinder the interoperability of blockchains
for example, is a smart contract
Modern blockchains such as Ethereum can advantages in terms of the safety, security company that uses a  High legal and regulatory uncertainty
run smart contracts that allow for the speed, accuracy and autonomy of combination of automated and manual remains, as governments are in the
automatic execution of transactions if pre- processes. Smart contracts techniques to audit the security of process of responding to developments
defined requirements are fulfilled. For further obviate data loss or smart contracts in order to create enabled by smart contracts
instance, a smart contract can perform error manipulation. These affordances can even more transparency in this field  Non-fungible tokens in particular have
checking for insurance claims and transfer be beneficial in parametric insurance been subject to price manipulation and
payments if conditions are met. money laundering schemes
 Smart contracts are giving rise to
DeFi applications. These applications
Risks
Entire organisations can be based on smart
contracts. A decentralised autonomous operate on an open, pseudonymous,  Ex-post correction of errors in smart
organisation (DAO) fully runs on a peer-to-peer basis and promise a contracts can be difficult or even
blockchain. The largest current faster, more flexible and less impossible. The flawless formulation
decentralised finance (DeFi) protocols are bureaucratic financial infrastructure. of smart contracts is critical. Unclear
managed as a DAO and are governed by The new technology attracts or unintended implications may arise
their members. institutional investors, which in turn if conditions stated in the smart
push more solutions to become contract are incomplete or ambiguous
Smart contracts enable the Internet of Value,
compliant. The possibility of losses
as digital tokens become non-reproducible
(in contrast to PDFs, for example,
Smart Contracts
Trial
Cyber & Crypto

Smart contract auditing to promote transparency

After the debacle and collapse of central crypto exchange FTX, decentralised finance is
Together with our partners Quantstamp and gaining more attention, as this space is not having single points of failure. DeFi is built
Chainproof, we are well positioned in order on top of blockchain technology, which enables the creation of decentralised
applications (dApps) that can facilitate various financial transactions, such as lending,
to seize the manifold opportunities that come borrowing, trading, and payments.
with the growth of smart contracts-based These dApps are typically built on blockchain networks like Ethereum, which provide the
necessary infrastructure for smart contracts. However, this comes with significant risk
business models. as hacks and exploits of smart contracts have occurred various times in the recent past.
Dirk Schaefer Therefore, various specialised auditing firms emerged (e.g., Quantstamp) to perform
Senior Underwriter (Digital Asset Protection) security and code reviews on smart contracts to identify vulnerabilities, security risks,
New Tech Underwriting and potential bugs in the code. Currently Munich Re is developing a smart contract
insurance cover in this new field.

Ethereum’s transition Risks Opportunities

to Proof of Stake  Staking is the act of depositing or  Staked assets are yielding rewards in
delegating deposits to active validator the form of digital assets which are
On September 15, 2022, the
nodes. The validator is responsible for received directly from the protocol.
Ethereum network (Eth)
Reduction of processing transactions or adding The rewards compensate for the effort
transitioned from a proof-of-work
(PoW) to a proof-of-stake 99.95% new blocks to the blockchain validators have to keep the blockchain
in total  Malicious behavior of a validator node running securely (e.g., maintain
(PoS) consensus mechanism.
energy use (e.g., through a technical failure) may hardware that runs execution clients).
With this step, the network
per transaction lead to a slashing and thus loss of Digital asset owner can stake their
took action against high energy
staked assets. Downtime of a validator assets to enhance their yield
resources consumption by
miners as network difficulty node may result in penalties by the  The insurance industry is currently
kept increasing. blockchain protocol developing staking event insurance
Eth PoW Eth PoS policies to protect asset owners against
loss of assets caused by slashing events
128
Privacy-enhancing Tech
Trial
Cyber & Crypto

Privacy-enhancing technologies (PETs) represent a wide range of Opportunities Risks

hardware or software solutions designed to maximise the social,  Third-party providers may handle data  PETs suffer from usability issues
scientific and commercial value of data, incorporating data related to insurance claims. Using PETs primarily with complex user interfaces,
protection principles through minimum data use, higher security and compliance with PETs principles leading to a poor user experience that
minimises access to customer data does not allow users to easily monitor all
and enhanced privacy.
 PETs promote the privacy of insurance aspects of their personal data
PETs enable businesses to leverage the increasing amount of data, with privacy as the customers by protecting customers’  Implementation and maintenance of
central tenet. They allow compliance with data protection regulations, including GDPR. financial details PETs are costly and require significant
Additionally, they enable data sharing with third-party entities without compromising computing resources and, as such, may
 Processing and validating insurance
data privacy. deter implementation
claims may require interaction with
Compliance with PETs also ensures the prevention of privacy breaches that adversely third-party organisations, including  While PETs offer many advantages,
affect brand reputation. PETs utilise various technologies, including different methods healthcare providers. The insurance however, they may lead to a false sense
of cryptographic encryption approaches (homomorphic, secure multi-party computation, sector can ensure customer privacy by of security, leading to higher data use in
zero-knowledge proofs, etc.), data masking techniques (e.g., obfuscation and utilising PETs. The use of machine breach of core PETs principles
pseudonymisation). learning and human recognition  Implementation of PETs by companies
Machine learning algorithms are also facilitated by PETs approaches. For example, systems may lead to the identification that enjoy large market shares may
synthetic data generation artificially creates data without compromising the real data. of previously unknown risks, resulting lead to monopolies, as they allow these
In addition, federated learning allows statistical learning without the need for data in better risk prediction, refined risk companies to analyse and use privately-
retention and centralisation. management and lower costs held data in customer devices to
 With the use of synthetic data, the increase revenue, for example, in the
insurance sector will not need to form of targeted ads in email accounts
capture a vast amount of insurance- and mobile devices
relevant real-world data to process a
claim. In addition, generating synthetic
data allows insurers to improve their
AI models without the need to access
privacy-restricted data, including
health records
129
Privacy-enhancing Tech
Trial
Cyber & Crypto

UK report: PETs for

better use of data
Source PETs are already revolutionising
Sharing and using data responsibly, lawfully, and securely royalsociety.org the way data is used, from
enabling greater cross-analysis
The role of PETs in data governance
between organisations to fueling AI
and collaborative analysis
in medical diagnostics.
Privacy-enhancing technologies The report argues that PETs could help unlock
“lifesaving” data without compromising privacy.
Encryption Pseudonymisation Anonymisation Better use of data could have significant public Prof. Alison Noble
Homomorphically Removing personal Irreversibly removing benefits – from cancer research to reaching University of Oxford, in report
encrypted data can information from the personal information net-zero carbon emissions. ‘From privacy to partnership’
be computed without dataset allowing data such that data owner
prior decryption. owners to remain cannot be re-identified.
unidentifiable.
Enveil Key Insight
Source
enveil.com
PETs enable companies to
Data summarisation Decentralised Differential privacy Unlock value with PETs leverage data while
Creating summarised learning Allows researchers and B2B startup Enveil is aiming to build complying with privacy
variants of data sets Allows sensitive data companies to collect and a new array of data products to regulations. The application of
that hide the actual to be offloaded to end- share aggregate ensure their users’ data privacy, using
information of individuals. information about users PETs can become a competitive
user devices where the homomorphic encryption and secure
processing can be done. while maintaining advantage for companies, as
multiparty computation. clients and other ecosystem
the privacy of
individual users. partners value secure and
private data handling.

130
Automated Compliance
Assess
Cyber & Crypto

As the significance and scale of data data breaches, insider attacks, or  Automated compliance solutions
increases incessantly, awareness of organisational abuses of collected data facilitate the design of new products, as
privacy and other compliance risks can occur. In addition to data, other the regulatory landscape can easily be
business processes are subject to screened and considered
has taken on a central role for
standards and rules. However,
businesses. Especially in the compliance with regulations tends
insurance sector, navigating the to be costly and difficult to enforce.
Risks
world of compliance is complex  Formalisation of regulations written in
Regulatory technology (RegTech) and
and resource-intensive. A rising natural language for automatic
other automated compliance solutions
number of software solutions enforcement is difficult to execute
aim to pre-empt harm to data and
promises so-called automated violation of regulations. They are also  Many data processing systems in use
compliance: systems that oversee, used in anti-money-laundering and today were designed and deployed
Know Your Customer (KYC) processes. before current privacy regulations were
monitor and ensure compliance
introduced and often hinder the
with regulatory requirements.
implementation of automated
The big data revolution is enabling new compliance solutions, for example by
business opportunities and helps create Opportunities offering no option to delete data
value for clients and companies alike.  In a complex regulation landscape,  Automatic compliance is further
Yet, given the great amounts of data, automated compliance systems complicated by frequently unstructured,
businesses are faced with compliance help teams to be in line with the richly formatted and lengthy documents
challenges. If proper handling fails, speed and breadth of regulatory
 Rather than full automation, many
change. Time savings due to
automated compliance tools provide
automation allow companies to
suggestions for human review
be proactive instead of reactive
and to allocate resources efficiently
 Automated compliance systems can
be used across lines of business
to mitigate risk and comply with
the numerous regulations
131
Automated Compliance
Assess
Cyber & Crypto

In an average week, how much time does your compliance team Key application areas for automated compliance solutions
spend tracking and analysing regulatory developments?
(in hours)

Less than 1 7% In contrast, 40% of Compliance Regulatory

Onboarding
globally systemic monitoring reporting
on average banks, said they spend
1 to 3 42% between
between 8 and 10 hours
4 to 7 31% 1 and 3 Based on: Thomson Reuters,
Financial
crime
Know-your-
hours Regulatory Intelligence Cost customer
defense
of Compliance 2022. Survey
8 to 10 20% of almost 500 practitioners
worldwide.

Drivers for modernising insurance compliance Fides Technology Key Insight

Source
fides.technology Compliance breaches pose an
Emerging technologies Regulatory pressures
e.g., RegTech and big e.g., new requirements One-stop-shop for corporate existential risk to a company’s brand,
data and analytics. and increased customer base, and ecosystem trust.
governance
inspections. Automated compliance solutions help
The startup aims to help companies insurance companies avoid these
fulfil their corporate governance duties outcomes. They also increase
without a million spreadsheets. The efficiencies in terms of time savings
Internal challenges company’s offering includes entity and error prevention. We expect
e.g., resource challenges, management, governance workflows, automated compliance systems to
inefficient coordination across and an audit trail of decisions. gain traction in the year ahead.
multiple jurisdictions.
132
Decentralised Data Economy
Assess
Cyber & Crypto

The decentralised data Data and other services are treated as Opportunities
valuable assets that can be traded and
economy is the cradle of exchanged transparently and securely
 Decentralised data centers allow for
better security and sovereignty of the
an open, transparent, using distributed ledger technology. data provided. Providers/owners of data
have and keep full authority and control
decentralised digital A decentralised data economy is a
over data
system in which companies can share
ecosystem where data and data between each other without losing  Businesses along the value chain get
liability, confidentiality and
services can be made it. And individuals share data on their
interoperability while using data
terms and choose to share it with others
available, collated, shared, in exchange for value. This removes  Less opportunities for data loss or
incorrect data to enter the workstream.
and monetised in an single points of failure and allows
By having decentralised data storage,
owners to retain control over their data.
environment of trust. Through the additional feature of
every entity has access to a real-time,
shared view of the data
blockchains, companies can offer
immutable data storage. Once
information is stored in the blockchain,
Risks
it cannot be changed or tampered with.  Decentralised blockchain systems often
prioritise security over performance.
By creating marketplaces, individuals, Adding members or actions to the
businesses and others can provide the network means it can be made safer,
data and use other data but not necessarily faster. Scaling up
pseudonymously, e.g., to improve their the network means slowing down
AI models. And data providers will have its performance
full transparency and control over how  Different types of identifier systems
their data is used, thus becoming with different implementations,
market participants. increasing the costs of development
and maintenance and making
integration between systems
much harder

133
Decentralised Data Economy
Assess
Cyber & Crypto

Example of decentral infrastructure layers 60% of all corporate

CAGR Data consumer Verifiable credential wallet Data provider data is stored in the cloud
+865% 2015 2025 Portal and API
(2022).
Statista

Access control (Compute-to-Data)

Comparing the year 2015 with

2025, there will be 175.8 Smart contracts

zettabytes of new data created.

Rethink Data Report Data Audit Trail (Blockchain)

Data is often referred

GEN-X network Use cases
Possible fields for data economy solutions in the future.
to as the fuel of the
Source
21st century. A
delta-dao.com
decentralised data
Pan-European network for the
economy could
Gaia-X Web3 ecosystem
GEN-X is a customisable blockchain
dissolve data silos
Mobility Analytics & AI Industry 4.0 Real Estate
built with Polygon Edge and leveraging while maintaining
Polygon Supernets. It enables the creation
of industry-specific data spaces that can
transparency and
Analyse mobility Usage of mass Help in Help in energy
be tailored to individual needs and use
infrastructure data with predictive consumption
control.
cases of consortia, SMEs or public
data. ML/AI. maintenance. prediction.
institutions.
134
Our Solutions
Cyber & Crypto
Munich Re
Cyber Solutions
Increasing our insureds’ resilience and improving the protection of
digital business models
While cyber risks pose ever greater threats to companies, individuals and society as a
whole, it is Munich Re’s aim to increase cyber resilience and responsiveness. Cyber
insurance solutions are fundamental for successful digitalization of the economy.
An adequate level of cyber security measures increases insureds’ resilience and is a
prerequisite for access to a sustainable cyber insurance market. In particular, systemic
risks and accumulation scenarios require a clearly defined risk appetite, clear wordings
and innovative solutions.
As a market leader, Munich Re offers capacity and services with comprehensive know-
how and supports stringent risk management. Together with our clients and partners,
we’re creating cyber solutions that go far beyond traditional risk transfer.

135
Our Solutions
Cyber & Crypto
Munich Re Munich Re Participants in the crypto economy can
Squalify Digital Assets and benefit from Munich Re’s strong financial
background, excellent reputation and 140
Speeding recovery after a Smart Contracts years of risk and insurance expertise.
catastrophic event The crypto economy includes
With rising cyber threats and a lack of new risks which require
methods to quantify the herewith related innovative solutions
cyber risks, Squalify provides a validated
The blockchain technology enabled the
and data-based approach to support
rise of digital assets and decentralised
companies in managing their cyber risks.
applications which triggered the
Squalify enables leaders of complex
development of a new ecosystem. Use
companies to professionally manage their
cases are of multiple nature with a
cyber risks with financial loss metrics. The
current focus on the disruption of the
top-down approach supports strategic
financial sector. With new solutions
decision-making on investments in
come new risks which require specific
information security.
know-how. Munich Re responds to
Using the long-standing and deep these new risks and works on solutions
expertise from Munich Re, Squalify can with the help of a newly formed Digital
grant unparalleled insights into cyber Asset Protection team combining the
risks. This service is delivered by separate crypto expertise available in the Munich
innovation venture to large partners who Re Group.
consult complex companies in strategic
A fundamental need of crypto
decisions in which Cyber risk plays a
stakeholders for security, reliability and
significant role. The vision of Squalify is
stability leads to extensive demand for
to become the global gold standard for
insurance capacity, primarily regarding
Cyber Risk Quantification.
the storage of digital assets, slashing
risk and functionality of smart contracts.
Redefining Industries
Technology impacts in 2023

50%
The ocean covers
two-thirds of the
world’s surface and absorbs
of all passenger
over 90% +12m vehicles sold worldwide will
of the excess heat. new jobs by 2030 be electric by 2032.
in the field of
Blue
Mitsubishi Heavy Economy.
Industries (MHI) is The concept of
using robots to
fight fires with full autonomy, 6x “Smart Cities” is spanning
and integrating a wide variety
delivering 4,000 liters more sustainable
of foam or water
seafood by of technologies – most of
2050.
per minute. them covered in this report
like AI, 5/6G, XR.

138
All trends 2023
Redefining Industries
The exploitation of previous technological advances have already demonstrated how businesses and Technologies for capturing carbon dioxide before it is released and building a resilience to survive
industries can be fundamentally revolutionised. Examples of industries that are currently being worldwide crises may offer solutions to some of today’s biggest problems.
redefined by technology include the renewable energy and manufacturing industry. Development of
renewable energy technologies such as solar, wind, and hydro power, along with energy storage Meanwhile, urban infrastructures are being redesigned to become “smart cities” and embrace
systems are changing the way energy is generated, distributed, and consumed. In 2023, we are autonomous mobility solutions. But it is not only cities that are becoming smarter. The agricultural
observing new trends including business resilience, blue economy and carbon capture. industry has also become a playground for digital solutions that enable greater efficiency and higher
outputs, and tech companies are now targeting the final frontiers, driving the “in-space economy”.

Adopt – Start initiatives in your unit Trial – Initiatives in affected units Assess – Evaluation needed Hold – Watch list
Robotic Revolution Autonomous Things Blue Economy
The application areas for robotics are broad, from Autonomous things are machines that are designed As the ocean is the largest common resource in the
manufacturing and IoT coupling, to household to work independently and without active human world, blue economy means economic growth based
helpers. Today there are different types from supervision. As AI improves, their functions are on the sustainable use of maritime resources.
industrial robots to sophisticated autonomous becoming more accurate and effective.
systems such as drones, self-driving cars and even
humanoid robots available. Business Resilience
Hydrogen Economy
Business resilience refers to a company’s ability to
The hydrogen economy refers to a proposed system
withstand and quickly recover from unexpected
in which hydrogen is used as a primary energy
disruptions such as natural disasters, cyberattacks,
carrier to power vehicles, buildings, and other
or economic downturns.
systems.

New Energy at Scale Carbon Capture

The production, storage, and incorporation of energy In the process of capturing carbon dioxide (CO2) the
into the electricity grid of sustainable energy at scale emissions are captured from power plants, industrial
is a challenge that is emerging with an increasing facilities and other sources before they are released into
share of and dependence on sustainable energy. the atmosphere.

Smart Cities In-Space Economy

A system of sensors, interconnected devices and Space-based services for objects and equipment in
services ranging from communications to healthcare, earth’s orbit, including orbital refueling, satellite
security and task automation from deployment and inspection.
private homes, public buildings and urban areas. new in 2023

139
Redefining Industries
Redefining Industries

Evolution over time

Adopt Trial Assess Hold

Start initiatives in your unit Initiatives in affected units Evaluation needed Watch list

Autonomous Things 2023 2022 2021 2020

Blue Economy 2023

Business Resilience 2023

Carbon Capture 2023

Hydrogen Economy 2023 2021

In-Space Economy 2023 2022 2021

New Energy at Scale 2023

Robotic Revolution 2023 2022 2021

Smart Cities 2023 2022

140
Impacts on the Value Chain
Redefining Industries

Explanation Insurance Value Chain

The illustration represents the impact the trend
fields have upon on the insurance value chain. Product Sales & Risk Customer Asset
With the dark boxes reflecting areas of high Design & Pricing Distribution Underwriting Policy Admin Management Engagement Claims Management
impact, and light boxes low impact.

Autonomous Things

Blue Economy

Business Resilience

Carbon Capture

Hydrogen Economy

In-Space Economy

New Energy at Scale

Robotic Revolution

Smart Cities

141
Robotic Revolution
Adopt
Redefining Industries

The robotic revolution is bringing significant changes to Opportunities

many industries. This includes increasing productivity,  In industrial and private contexts,
reducing labour costs and improving safety. In addition, it robots require insurance coverage –
will facilitate the shift to more sustainable and efficient opportunities for insurers to design
robot insurance products
production processes.
 Robots can serve as remote claim
The robotic revolution refers to the increasing use of robots and automation in various
industries, such as manufacturing, healthcare, transportation, and many more. The
detectors for insurers Risks
robotic revolution is driven by advances in technology such as artificial intelligence,  Robots replacing a human workforce
 Cyber security risks: As robots become
machine learning, and sensor technology, which have led to the development of more reduce the risks of severe accidents in
more connected to the internet and
advanced and capable robots that are able to perform a wider range of tasks. tasks posing a high risk to life (such as
operate in more sensitive environments,
bomb defusing and window cleaning
Robotic technology has developed rapidly over the years, and there are many they can become vulnerable to
for skyscrapers)
different types of robots available today, from simple industrial robots to cyber-attacks, leading to security
sophisticated autonomous systems such as drones, self-driving cars and even  Robots can assist human workers in breaches and loss of data
humanoid robots. Mitsubishi Heavy Industries (MHI) is using robots to fight areas such as precision, strength and
 Modeling risks: It is strongly related
fires with full autonomy, delivering 4,000 liters of foam or water per minute. endurance, allowing for more efficient
to the question of the insurability of
RoboBees, tiny Harvard-developed bots, use electrostatic adhesion to “perch” and safer work environments
algorithms, which represent the
on walls and ceilings to evaluate structural damage after earthquakes.
 Claims savings: Robots can help to fundamentals for robot actions
The robotic revolution is being driven by several key technologies, including reduce labour costs and increase
advancements in hardware and sensors, batteries and electric motors. Some of those  Safety risks: Robots can malfunction, or
efficiency, which can lead to savings in
technologies have undergone a major transformation due to advanced deep learning in be hacked, and as they are being used in
claims expenses
the recent years, enabling robots to perform a wider range of tasks and to work more more critical industries like healthcare
effectively with humans. or transportation, these risks can
become critical for human lives
Collaborative robots, also known as cobots, are also a driving force behind the revolution,
as they are designed to work alongside humans as robotic colleagues. However, this also
raises concerns about job displacement, as well as questions about the ethical and legal
implications of autonomous systems.
142
Robotic Revolution
Adopt
Redefining Industries

Amazon's Robot Serving Robot

Key Insight
Sparrow Plato
Industry Industry
Improved sensor technologies
Logistics Tourism/Hospitality
and AI make robots more capable
Source Source of interacting with humans and the
TechRadar United Robotics Group physical world around them.

Amazon’s warehouse robot can identify United Robotics has developed a robot Robotic systems and their potential
around 65% of its product inventory for hotels and restaurants. Plato is able to for insurance coverage and services
without human assistance thanks to an bring food to the table or bring the dishes therefore need to be evaluated.
integrated AI and computer vision. back to the kitchen. As a result, the
employees have more time for the clients.
143
Autonomous Things
Trial
Redefining Industries

Autonomous Things, also Opportunities

known as “smart” or “self-  Autonomous Things that “share”
governing” things, refer to their workplace with humans
(e.g., in warehouses), require
physical devices or systems innovative liability insurance solutions
that can perform tasks without  Autonomous Things allow insurance
human intervention. As AI companies to improve their under-
improves, their functions are writing of property and casualty risks
through upfront inspections. Moreover,
becoming more accurate periodic inspections indicate whether
and effective. the current coverage is still appropriate

AI is propelling advancements in new

intelligent objects like drones and self- Risks
driving cars. The technologies are evolving
 Safety concerns: Autonomous
rapidly, thanks to new machine learning
Things share an operating environment
models and algorithms, allowing them to
with humans. e.g., self-driving cars, may
function without human interaction for a
not always make the same decisions as
set period.
human drivers, potentially resulting in
These smart devices gather information, accidents or injuries
make decisions, and carry out actions in
 Privacy and security risks: Autonomous
the physical world using a combination of
Things are often connected to the
sensors, artificial intelligence, and
internet, which makes them vulnerable
connectivity. Autonomous Things aims to
to hacking or other cyber attacks. This
make devices and systems more efficient,
could lead to sensitive data being stolen
dependable, and convenient by eliminating
or systems being compromised
the need for human intervention and
allowing them to adapt to changing
conditions.

144
Autonomous Things
Trial
Redefining Industries

Amazons Warehouse Robot Proteus Waymo Self-driving Taxi

Autonomous Proteus is Amazons first fully-autonomous robot. The bot One Waymo One is a self-driving taxi located in

Robot is used in Amazons warehouses and lifts heavy products

from one point to another. Industry
San Francisco and Metro Phoenix. According to Waymo
you can call the autonomous vehicle to your location via
According to Amazon Potreus was built to work Mobility app and it will drive you to any point in the city.
Industry
autonomously and around the employees. The robot is Additionally, Waymo is using their taxis as weather
Logistics able to stop on his own if it crosses paths with a human Source gauges. According to the company the car sensors are
Source worker. This is possible due to Amazon Notes, the Waymo able to create real time weather maps. This can be
company's proprietary, advanced safety, perception, and helpful to test the limitations of Waymo Ones driving
TechRadar navigation technology. skills in simulations as self-driving vehicles normally
having problems with weather effects.

Key Insight

The concept of Autonomous Things is originated in the domain of

drones, driverless vehicles and vessels, and is now being adopted in
various other domains such as health, logistics and traffic.
Based on the maturity of the technology and adaption of the regulatory
and legal environment, various new needs for risk protection arise.

145
Hydrogen Economy
Trial
Redefining Industries

The hydrogen economy refers to a system design in Opportunities

which hydrogen is used as a primary energy carrier to The hydrogen economy has the
power vehicles, buildings, and other systems. potential to reshape the way we
produce transport and consume energy
Hydrogen would be produced by using renewable energy sources such as solar or requiring new services and insurance
wind power and then used in fuel cells to generate electricity in this system. solutions across those chains addressing
Because water is produced as a byproduct of this process, hydrogen is a clean and the needs of big players and start-ups
sustainable energy source with no emissions. The hydrogen economy is viewed as (owning core technology).
a possible solution for reducing reliance on fossil fuels and mitigating the effects of
climate change. However, it is still in its infancy with many technical and economic
hurdles to overcome before it can be widely adopted.

The hydrogen economy could be a

potential solution for decarbonising
transportation and storing energy from Risks
intermittent renewable energy sources,
 There are several risks associated
hydrogen fuel cell technology is already
with the implementation of a hydrogen
used in some vehicles and stationary
economy, including safety, storage and
energy storage applications.
distribution, environmental impact,
However, before it can be fully realised, lack of regulation as well as technical
infrastructure for hydrogen production, challenges
storage, distribution, and fueling must be
 Those risks will require careful
developed and widely deployed.
modelling and design of new risk
coverage solutions

146
Hydrogen Economy
Trial
Redefining Industries

A fossil-free future in the ITM's Proton Exchange

Key Insight
steel industry Membrane Technology
Industry Industry
Tangible use cases will drive the
Steel Engineering and Industrial speed of application of hydrogen
Source Source economy
HYBRIT London Stock Exchange Hydrogen has the potential to play a
significant role in the transition to
In the steel industry, companies such as ITM's first-of-a-kind technology to
clean energy. Its successful adoption
SSAB, LKAB, and Vattenfall have formed produce green hydrogen gas via
and implementation will depend on
a joint venture called HYBRIT, which aims electrolysis is seeing a new phase of
progress in technology,
to use hydrogen instead of coking coal to growth begin to create efficiency
infrastructure, and policy support.
produce steel and ultimately achieve a and reliability.
fossil-free steel-making process.
147
New Energy at Scale
Trial
Redefining Industries

The production, storage, and  Geothermal: Using heat from the  If a group of prosumers are providing
Earth’s core to generate electricity by energy to an industrial customer, this
incorporation of energy into heating a fluid to create steam, which industrial company needs insurance
the electricity grid of then turns a turbine to produce against the prosumers not being able
electricity to provide enough energy
sustainable energy at scale is
 The whole energy ecosystem
a challenge that is emerging All of those methods have their own
advantages and disadvantages, and the becomes more complex because of
with an increasing share of suitability of each method depends on the integration of different sectors like
and dependence on the specific local conditions. electricity, heat and transportation. A
stable electricity supply will become
sustainable energy. An emerging approach to producing
the key part of our economy but needs
sustainable energy at scale is the
There are several approaches to produce risk management
Cellular Smart Grid Platform (CSGriP).
sustainable energy at scale, e.g., This platform enhances the grid concept
 Solar power: Using photovoltaic cells to through the better utilisation of energy Risks
convert sunlight into electricity, large- distribution networks. The goal is to  Technology risks are unknown,
scale solar power plants build an intrinsically stable grid unit that both in early stages and long-term
runs without the need for a far-reaching impacts on the environment (e.g.,
 Wind power: Using wind turbines to
and complex infrastructure. decommissioning and recycling)
convert the kinetic energy of the wind
into electricity, large wind farms  The whole ecosystem will be much
 Hydroelectric power: Using the energy more complex to model and manage,
of falling water to turn a turbine and compared to today
generate electricity, hydroelectric Opportunities
power plants
 New concepts for performance
 Biomass: Using organic matter, such as guarantees. The market will shift to
wood or agricultural waste, to generate a prosumer (producer and consumer
electricity through combustion or at the same time) market with millions
fermentation of electricity suppliers
148
New Energy at Scale
Trial
Redefining Industries

AREMI Program Key Insight

Industry Source
Manufacturing PV Tech New energy at scale spans a range of new technologies which will
not only need to work together but also integrate with legacy supply
The Africa Renewable Energy chains, for example. This could imply secondary risk (e.g., on the
Manufacturing Initiative (AREMI) supply chain), which in turn could impact existing insurance
launched with the goal to expand the products and coverage. However, on the positive side it could
capabilities of renewable energy introduce opportunities for new data-driven insurance solutions.
manufacturing in Africa. The goal of
the initiative is two-fold: reduce the
dependence upon fossil fuels and
create significant new job opportunities.

Amazon starts trading

renewable energy
Industry Source
Energy Carbon Credits

Based on green light from the Indian

authorities, Amazon will start
trading renewable energy resources
and developing new wind and solar
energy projects in India.

149
Smart Cities
Trial
Redefining Industries

A smart city is an urban area There are a number of countries globally  Smart devices can track and learn
who are designing and building smart, individuals’ daily routines and therefore
that utilises technology to technology advanced cities or islands. allow people to monitor their elderly
enhance the quality of life Current well known smart city projects relatives without invading their privacy
include Neom/The Line (Saudia Arabia),
for its residents. It also uses  Insurers benefit from smart devices
Energy Island in Denmark, two in Korea because they lower an owner’s
technology to improve the and even in Germany (Essen). risk profile, deliver relevant data for
use of resources and There are as well projects on the way
underwriting purposes, and improve
the efficiency of claims
increase the efficiency of combining technologies for a smart use settlement processes
case like airports implementing
urban services. biometric recognition systems for  Smart technology will lead to a wide
security purposes and cities are range of new products (e.g., flexible
Smart cities use a combination of digital, household insurance based upon
optimising traffic in real time,
information and communication shared data)
using smart traffic lights.
technologies to better manage assets and
resources, reduce waste and costs, and
enhance communication and decision-
Risks
making processes. This can include the  IoT systems create a back door for
use of sensors, Internet of Things (IoT) hackers to steal data and control
devices, big data analytics and other important functions such as opening
technologies to gather and analyse data. the front door, adjusting the heating
system, etc.
 Only a few private technology
Opportunities companies are likely to serve the smart
space market in the future, and these
 The installation of IoT technology may must therefore be strictly controlled to
substantially increase property value avoid data misuse, such as monitoring
and provide damage reduction behavior, purchase decisions, etc.
by prevention

150
Smart Cities
Trial
Redefining Industries

Building Data insights based on A new Blueprint for cities build from ground-up
Data geospatial data analysis kind of city Neom has announced a floating industrial district to be
Credium is a german AI-company that is focused on implemented in the emerging smart city in Saudi
Insights delivering building analytics. Credium provides a variety Industry
Arabia. Oxagon is to host a hyperscale data centre in
of real estate data such as building geometry or the size partnership with Fas Energy, which will power
Industry of living space for specific buildings via a web interface
Various the hexagonal industrial district. A port and logistics
centre will also be integrated. Oxagon will particularly
Various (API) on the basis of an address only. Source
promote hydrogen projects and house a modular building
Source
According to the company there are no consistently high- Neom/Oxagon construction factory. According to the initiator
quality building and property data available. Collecting
Credium these data is usually very time consuming and expensive
Prince Salman, Oxagon will make a major contribution to
regional trade and support new global trade flows. The
because it must be selected by hand. Crediums goal is it
construction of "The Line" started in 2022.
to have a solution that automates the process.

Smart Sunderland plan to become a smart city Key Insight

Sunderland The Sunderland City Council, in partnership with BAI
Communications, is setting their goal to become UK's
Wide range of technologies orchestrated
most advanced smart city.
Industry
From providing a 5G network in the city center to smart The concept of “Smart Cities” is spanning and integrating a wide variety of
Various education or smart bins, smart technologies can be technologies - most of them covered in this report like AI, 5/6G, XR, sustainable
Source beneficial to enhance the everyday life. Industry in the energy, cyber, robotics/automation - offering opportunities for new and
local area, including manufacturing, logistics and social adapted business. This may include “pay for use” insurance models,
BAI performance guarantees and embedded insurance services. Depending on
care can also profit from the roll out of digital services.
Communications the investment appetite, start piloting.

151
Blue Economy
Assess
Redefining Industries

The ocean is a life-giving Still, the term blue economy is new to such as mangroves, coral reefs, or even
many of us. Per definition, it translates to Marine Protected Areas
resource and offers vast economic growth based on the
economic potential. The sustainable use of maritime resources.  For the marine lines of business,
insurers can use the blue economy’s
concept of Blue Economy aims The blue economy comprehends not momentum to decarbonise their
to embrace these economic only crucial industries such as food, portfolio through offshore wind farms,
energy, and transport but also wave and tidal energy, and even floating
opportunities while paying
recreational sectors like tourism, solar panels
adequate attention to the and complex industries, such as
 Scalability opportunities for aquaculture
future health and productivity biotechnology and carbon storage.
insurers thanks to the increasing
However, for this economic model to
of these resources. prosper, capacity building and the demand and consequent supply
participation of the financial sector for seafood
The ocean covers two-thirds of the
are key components.
world’s surface and absorbs over 90% of
the excess heat and more than 25% of the Moreover, as the ocean is the largest
Risks
additional carbon dioxide emissions common resource in the world, effective  With the inclusion of the shipping
generated by human activity around the coordination among different sectors, industry in the European Emission
world. It is the main source of protein stakeholders, and governments are Trading System, new liabilities arise for
for over 3 billion people and has an fundamental to ensure the development ship owners or managers
of a sustainable, inclusive, and equitable
estimated net asset value of  Reputational risks might arise when
blue economy.
US$ 24 trillion. insuring unsustainable and extractive
economic activities, such as deep-sea
mining
Opportunities  A better understanding of the ocean
ecosystem is needed for NatCat
 The increasing awareness, importance modeling and accumulation planning
and valuation of ecosystem services
allows insurers to offer tailor-made
solutions for crucial natural capital,

152
Blue Economy
Assess
Redefining Industries

Benefit-cost ratios and net benefit of ocean-based Floating Platforms

interventions1 Industry
by 2050 Energy
Offshore wind Source
12:1 US$3.5tn
energy production SINN Power
Ocean-based The solution from SINN Power includes
10:1 US$6.7tn
food production floating platforms for lakes, lagoons and
Decarbonisation of the ocean. The so called SLake and
4:1 US$5.1tn SLagoon platforms provide renewable
International shipping
solar energy whereas the SOcean hybrid
Conservation and
restoration of mangroves
3:1 US$0.2tn platform can transform solar, wind and
wave energy.

Economic, social, and environmental benefits Washing the Oceans

A sustainable blue economy can result in: Industry
Food
1/5 Source
of GHG emission Origin by Ocean
reductions Origin by Ocean’s goal is to establish a
12 Million 6× more 40× more needed to keep business that uses the harvesting and
new jobs by sustainable renewable the world within processing of algae as its foundation.
According to the company the algae is
2030 seafood by energy by 1.5°C turned into ingredients that can be used
2050 2050
for production of consumer goods like
food or cosmetics.
153
Business Resilience
Assess
Redefining Industries

Business resilience refers to a company’s ability to Opportunities

withstand and quickly recover from unexpected  Business resilience not only helps
disruptions like natural disasters or economic organisations to manage and
downturns – and lately to an increasing extent from recover from unexpected
disruptions, but it also provides
technology-related issues such as cyberattacks or IT opportunities for improvement and
supply chain problems. growth in the insurance sector
Business resilience helps organisations to be able to maintain and recover  By taking advantage of these
their core functions and critical operations in the face of unexpected opportunities, organisations can
challenges and events. This includes having plans and procedures in place to not only survive unexpected
minimise the impact of such events, as well as the ability to quickly adapt and disruptions, but also emerge
continue operations. stronger and more successful.
Insurance can design new
Technologies impact the meaning of business resilience. On the one hand side, services and offerings to act as
they can strengthen a company’s and society’s ability to cope with adverse risk management partners and
incidents, such rising sea levels or extreme weather conditions. In its 2022 provide overall coverage as well
Climate Change Report, the Intergovernmental Panel on Climate Change
points out the role of technological innovation for climate resilience. Further,
data sciences have enabled sophisticated prediction capabilities, helping Risks
companies prepare for and manage disruptive events.  High accumulation risks for
On the other hand, technology adoption and technology-related issues insurers, especially for events that
challenge business resilience. For instance, IT-related productivity may break have a cumulative impact e.g.,
down during catastrophic events, or chip shortages may impede production. across geographies and industries.
Companies need to put in place a holistic digitisation strategy that offers This requires new models
safety and contingency plans for its hard- and software operations as well as and risk simulations
for its stakeholder relationships. A sense of agility can further help companies
react quickly and collaboratively.
154
Business Resilience
Assess
Redefining Industries

Supply Supply Chain Risk Management Continuity Storage and Data Protection
Chain Sphera is specialised in analysing risk management in
supply chains. This is accomplished by using artificial
Software Continuity Software provides their customers with
protection against cyber security threats. The company
intelligence, big data and machine learning to identify focuses on protecting the storage and backup systems
Industry Industry
supply chain risks in real time. on the ground as well as in the cloud.
Various In addition, according to the company, the extent of the
Various For their work, the company was recognised as a Sample
Source damage will be evaluated, supplies will be provided and Source Vendor by Gartner in the Hype Cycle for Storage and
appropriate measures will be initiated to protect Data Protection Technologies 2022.
Sphera Continuity
the brand.
Software

Key Insight

Currently we are experiencing a level of disruption and business risk not seen in
generations. Some companies freeze and fail, while others innovate, advance,
and even thrive. Business Resilience improves a company’s ability to handle
unforeseen events and crises by preparing for different scenarios. It also
increases the company’s reaction capacity and allows the company to learn and
grow from crises. Business resilience capabilities will make the difference for
long term success.

155
Carbon Capture
Assess
Redefining Industries

Carbon capture refers to the Opportunities

process of capturing carbon  Reducing greenhouse gas emissions:
dioxide (CO2) emissions from Carbon capture can significantly reduce
the amount of CO2 emissions from
power plants, industrial power plants and other industrial
facilities, and other sources facilities, which can help to slow the
pace of climate change
before they could be released
 Facilitating compliance with
into the atmosphere. environmental regulations:
The reduction of carbon emissions is one Carbon capture technology can help
of the key strategies for tackling global organisations to comply with regulations
warming. The process of carbon capture related to greenhouse gas emissions and
and storage (CCS) starts with the capture can also support organisations’ efforts
of CO2, optimally directly where it is to reach net-zero emissions targets
produced, such as at power generation  Supporting Sustainable Development
and industrial grounds. The captured CO2 Goals (SDG): Carbon capture and
is then either transported and stored deep storage can contribute to a number of
underground or, alternatively, it may be the SDGs, such as SDG 7 (affordable
used in industrial processes or potentially and clean energy), SDG 9 (industry,
converted into other useful products. innovation and infrastructure) and
The carbon capture technology is still SDG 13 (Climate Action)
relatively new and expensive, and its large-
scale implementation is currently limited. Risks
However, it has the potential to signifi-
 Carbon capture technology is relatively
cantly reduce greenhouse gas emissions
new and still has several risks and
and mitigate the effects of climate change.
challenges associated with it, such as
high cost, storage and transportation,
safety and environmental risks

156
Carbon Capture
Assess
Redefining Industries

Shoe made from Carbon capture

carbon emissions in India
Industry Industry
Manufacturing Energy
Source Source
LanzaTech Carbon Clean
A shoe made from carbon emissions Carbon Clean begins carbon
was presented by Swiss sports brand capturing at NTPC's coal-fired
On. Lanzatech develops a technology power plant in Madhya Pradesh,
that captures carbon emissions from India. According to the two
industrial processes and converts companies the carbon capture
them into chemicals and fuels. In this plant is intended to capture 20
scenario the carbon emissions were tons CO2 per day and will later Key Insight
converted to a new foam material be combined with hydrogen to
used for the shoe. produce 10 tons of methanol
Carbon capture initiatives will provide environment, economic
per day.
and social benefits bringing positive impact locally and
globally. Insurance could play a vital by securing effective
insurance solutions and protecting the financial stability of a
carbon capture initiatives through risk mitigation techniques
during construction, operation and decommissioning phases.

157
In-Space Economy
Assess
Redefining Industries

Space-based services Given the ever-growing volume of data

sent through orbit via satellites, in-
Opportunities
for objects and space services for orbital objects are  Enables ‘second life’ for satellite assets,
as satellites can be refueled and
evolving into a whole economic system.
equipment in earth’s The market is currently dominated by updated in space for different uses
private companies such as Elon Musk’s
orbit, including orbital SpaceX and by public institutions and
 Reduces risk of collision as defunct
satellites can be repaired or de-orbited
refueling, satellite governments that are opening up the
market for (private) companies. This is  Creates new opportunities for risk
transfer solutions guaranteeing success
deployment and driving rapid growth in the sector
of servicing operations
and opening a broad range of business
inspection. opportunities. For example, the  Enables better understanding of failure
manufacture of reusable rockets or modes if a failed satellite can be directly
satellites that can be refueled or inspected from space
repaired, on-orbit servicing for satellites
via robotics and/or AI, satellite Risks
communications for IoT, and low-cost
miniaturised satellites able to collect  In-space servicing involves close contact
data or monitor larger satellites. For this between expensive assets, which
adds new hardware risks
reason, the sector will see increasing
commercialisation as companies start  Unproven technology is likely to suffer
offering satellite broadband, (private) failures at first until the technology and
space tourism for everyone, and high- market matures
speed delivery of packages. However,  Significantly increased governmental
there will be a need for greater and commercial space activities will
governance in terms of “orbit entail, that more and more inactive
ground control”. satellites and other uncontrolled objects
remain in orbit and thus the collision risk
of space debris will be increased

158
In-Space Economy
Assess
Redefining Industries

In-space The start up Orbit Fab plans to deploy a

refueling refueling shuttle and a refueling depot for GEO
spacecrafts. Satellites can be refueled with
Industry
hydrazine at the refueling station in space or a
Aerospace
shuttle can navigate to them. With a price tag of
Source US$ 20 million, the project is scheduled to
Spacenews launch in 2025.

Microgravity Space factory startup Varda is planning a Key Insight

manufacture new method to manufacture tools or other
products such as pharmaceuticals in The in-space economy is expanding and becoming increasingly global,
Industry
microgravity. To accomplish this, Varda driven by the development of ever-more governmental and private
Aerospace sector initiatives around the world.
wants to launch a manufacturing satellite into
Source In the future space activities will be more and more commercialised and the
space. To advance this project, the company
number of stakeholders will considerably increase compared to today. Through
CNBC partnered with NASA. technological progress and new business models such as ridesharing for small
sat operators there will be a high number of satellites in orbit. While this
development drives innovation, it acts also as a tremendous risk for the
sustainability of the space industry.

159
Our Solutions
Redefining Industries
Munich Re
Location Risk Intelligence Platform
Assessment of physical risks associated with climate change
and natural hazards
The cloud-based platform provides organisations with a comprehensive solution to
analyse physical climate and natural hazard risks globally for multiple perils, from single
assets to an entire portfolio. By entering dedicated locations, users can access risk and
hazard assessments in the form of risk scores for a variety of natural hazards and/or
climate change risks, visualize them and save or download them as comprehensive
reports.

Benefits at a glance
 Covers a comprehensive set of 12 current and 7 future climate hazards
for physical risk identification
 Future climate hazards are built on IPCC’s proposed RCP scenarios
(2.6, 4.5 and 8.5)
 On-demand version, users without a contractual commitment can purchase
risk assessments for individual locations

160
Our Solutions
Redefining Industries
Munich Re Munich Re
Hydrogen Equipment as a Service
Technology-driven solutions Leveraging industrial machine IoT
enable the energy transition Hydrogen from electrolysis is data to optimally transfer and
the key element for reversing a manage asset ownership risks
Hydrogen produced from renewable
formula which had been followed
energy sources is one of the most EaaS packages risk solutions, asset
important alternatives to fossil fuels and for decades.
investment and digital services from the
will become the game changer for the Munich Re Group, together with partners’
energy transition. Production use and Renewable energy can industrial equipment and technical services,
conversion are evolving quickly. now be used not only into an integrated service offering in which
Munich Re’s Green Tech Solutions helps customers only pay for the use or output.
pioneers in the hydrogen domain to
24/7 but all year long.
Technology and data analytics are key
leverage the opportunities and cushion
the risks involved in hydrogen production
To be sustainable this enablers for automating operations and
admin processes, creating digital
with the focus on electrolysers and other emerging technology interfaces with customers and partners,
areas of the hydrogen value chain. We
partner with forerunners in those new
needs political support and optimally managing risks.
Benefits at a glance
technologies, securing predictability of and bespoke risk
cash flows and the investments in the  Covers a comprehensive set of
long run.
management. 12 current and 7 future climate hazards
for physical risk identification
Benefits at a glance  Future climate hazards are built on
 Tackling technology and financial risks IPCC’s proposed RCP scenarios
of scaling emerging tech (2.6, 4.5 and 8.5)
Patrick Hinze
 Managing accumulation and  On-demand version, users without a
counterparty risks Head of contractual commitment can purchase
Hydrogen risk assessments for individual locations

161
Our Solutions
Redefining Industries
Munich Re
Infrastructure Risk Profiler
Holistic risk assessment for infrastructure investments
Proper analysis of risks typically associated with infrastructure requires comprehensive
expertise drawn from diverse fields.
The extensive IRP analysis covers all risk factors pertinent to infrastructure projects:
macroeconomics, technology, natural hazards, project execution and operation,
environmental impact, microeconomics, insurance, as well as ESG. The approach
considers and weighs relevant risks individually and holistically.
Benefits at a glance
 Holistic, objective and transparent perspective
 Solid basis for an informed investment decision to better secure a return
on investments
 Thorough analysis within up to 4 weeks
 Comparability of different infrastructure projects that match clients’
individual appetite for risk

Our competence, your confidence.

Franz Vogt

Senior Project Manager

Infrastructure Risk Profiler

162
Our Solutions
Redefining Industries
Munich Re classification of EVs according to their Munich Re
EVA – Electric Vehicles specifications. Leveraging Munich Re’s IoT Solutions
global exposure and advanced ML
Analytics techniques, we identify the vehicles’ IoT-based business models shift risks from clients to providers
The next level of mobility intrinsic risks and rank them accordingly. Medium-sized machinery and plant companies are changing from analog product
Global Consulting developed a solution to The assessment of (new) vehicles is easier sellers to digital solution providers. Promising business models based on IoT
enable insurers who want to underwrite thanks to EVA, allowing insurers to better applications or X-as-a-Service models are creating new and stable revenue streams and
and price EVs with the appropriate estimate the risk, define a clear risk deeply networked customer relationships. For providers, however, this also means that
confidence. Our sound statistical analysis appetite, and underwrite new risks with they have to take increasingly more of their end customers’ risks onto their own
enables the identification and more confidence. balance sheets.
Those who outsource the defined performance guarantees to a risk carrier like Munich
Re relieve their own balance sheet. With Munich Re’s seal of approval, they build up
Differentiation between non-plug-in electric vehicles and plug-in electric vehicles
trust in the performance of their solutions and can boost their sales.
ICE & Non-Plug-in Plug-in
Electric Vehicles Electric Vehicles (EV)

Petrol and Diesel Vehicles (ICE) Battery Electric Vehicles (BEV)

Biofuels/Natural-Gas Vehicles 100% electric powertrain with battery
(minor market share). range up to 800km. Battery Charged
via plug-in to power source.

Hybrid Electric Vehicles (HEV) Plug-in Hybrid Electric Vehicles (PHEV)

No Plug-in: Combination of conventional Combination of electric and ICE
ICE system with electric propulsion powertrain with up to 80km electric
system (mild- or full-hybrid). range. Battery charged via plug-in to
power source.

Our focus
163
Our Solutions
Redefining Industries
Munich Re
Epidemic Risk Solutions
Financial protection from catastrophic disease outbreaks – building a more resilient future
Epidemic and pandemic outbreaks are becoming more frequent and severe. They cause serious economic impacts on all business
sectors, economies and societies. In the event of an outbreak, rapid action is required to protect the business and all stakeholders.
This solution provides a cash injection based on clear triggers and expedited claims settlements to cover all your needs, expected
and unexpected.
Benefits at a glance
 Simple and transparent parametric triggers
 Expedited claims settlement to quickly access funds
 Meet ESG obligations by protecting employees and reassuring external stakeholders
 Expert knowledge and advice: access to epidemiological and specialised risk management expertise
Broader context: To make pandemic risk insurable, redefining the risk transfer value chain towards an Epidemic Risk Markets
Platform is necessary to combine different forms of risk financing from different sources of capital. APEC Finance Ministers have
endorsed the implementation of the Epidemic Risk Market Platform in Public-Private Partnership.

How can we increase resilience against pandemics?

By spreading the risks across insurance and capital markets.
Dr. Gunther Kraut

Head of Epidemic
Risk Solutions

164
Regional Pages
United States perspective
North America

How is the market driving

technology adoption?
 Artificial Intelligence (AI) and Internet of Things (IoT) are two examples of Norman Thoms
new technologies that are creating opportunities that improve operations
and create a more personalised service. Insurance companies are adopting VP IoT Technologies
these new technologies to better analyse risk (IoT) and to automate specific Hartford-HSB
and unique types of insurance processing (AI). These new technologies will
ultimately improve the customers experiences
 Despite digital and direct market offerings being developed, long-
forecasted disintermediation of sales process has not materialised Online acquisition strategies are yet to reduce
due to product complexity and regulation
huge customer acquisition costs, currently at
 Digitalisation mainly enables ‘research online – buy offline’ trend. The main a large portion of premium.
efficiency gain on insurer side is administration cost reduction
 Regulation is a major driver for lack of new product development. This is Now, tech is being used to improve
caused by state-fragmented regulatory landscape. Reg-tech can potentially
fill this gap
productivity of agents and brokers
 Technology investment into parametrisation of risk and risk prevention, on the one side and product
and insuring digital assets is likely to be rewarded. The reason for this is
capability for scaled solutions leveraging the technology investment
development on the other.

166
United States perspective
North America

Assess Insure
Use Case
Continuous risk Adaptive cyber
Cowbell is a provider of adaptive cyber insurance for SMEs. assessment insurance
Founded in 2019, the company has built a self-service, automated
journey for brokers to offer cyber coverage to their clients. The
resulting efficiencies, combined with best-in-class underwriting,
have attracted more than 21,000 brokers.
Respond Improve
Cowbell is able to automatically underwrite many organisations by
cataloging risk data across a large part of the U.S. SME market. This In-house cyber Add-on services
risk pool is used to instantly provide bindable quotes while policies are claim experts & risk engineering
issued in under 5 minutes. More complex cases are augmented with
human underwriting.
Agent network growth of Cowbell
Throughout the life cycle of their policy, insureds are advised on how June 2020 – December 2022
to reduce cyber risk and optimize premium, thereby boosting
customer retention for brokers.
20,000

10,000

2020 2021 2022

167
Asia-Pacific perspective
Asia

How is the market driving

technology adoption?
 This is an agent-driven market and significant investments have
Nick van der Nest
been made into these traditional distribution channels as they are Chief Propositions and Innovation Officer
deemed very effective
Asia Pacific, Middle East and Africa
 Established and successful insurers are investing in digitising and
connecting their processes to generate data from which AI models
can be created
 Agent productivity improvements are being achieved through
the use of AI models, which enable underwriting decisions to
AI is being used to supercharge
be predicted without human involvement automated underwriting engines
 Due to the popularity of bancassurance in these markets, that agents rely on.
augmentation of the onboarding journey is also a priority
 Technology solutions in China are very different to the rest of the
Asia-Pacific markets due to vastly different privacy regulations
 For example, the most popular social media platform, WeChat,
is used for insurance sales, regular insurance payments and
customer services
168
Asia-Pacific perspective
Asia

Use Case
Munich Re‘s Allfinanz underwriting automation system
has been maturing for over 2 decades.
There are over 50 installations of this software in the Asia-Pacific
region and rich data is being extracted by users of this software to
gain insights into augmentation of traditional underwriting with AI
models.

Accept
Automated 65%
Load Streamlined
New approach process for
with AI Model Human Decline 35% agent and client

Insurance Agent e.application Insurer's

buyer completed system
Automated Accept 20%
Old approach Not optimised
Human Decline 80%
Source: Improvement in one project Munich Re was involved in 169
UK perspective
UK

How is the market driving

technology adoption? Simba Zvauya
 The UK is a mature insurance market characterised by CUO – Portfolio Solutions
high competition among incumbent insurers, aiming to Munich Re Specialty Group
break out into more profitable offerings. Aggregators and
London – United Kingdom
comparison websites have eroded value from traditional
product lines (e.g., motor insurance)
 Despite the market being mature inefficiencies increase
as the size of the risk and complexity of the product
increases. Using the learnings from the digitalisation Insurers are looking to gain margin in the
within personal lines, insurers are now in the process of
investing in the digitalization of larger more complex P&C market by leveraging technologies
risks. This is particularly evident in the SME segment
which enable intermediaries to advise on
 Insurers, brokers and software houses are focusing on
developing digital distribution ecosystems that accelerate a wide variety of complex products
underwriting, pricing and policy renewals, including
automation of some human decision-making whilst increasing efficiency.

170
UK perspective
UK

Use Case Brokers using Acturis …

The UK market has seen software houses that provide broker solutions investing Brokers have realised that the increased penetration of e-trade has allowed them to
into creating end-to-end ecosystems that integrate brokers and insurers. These improve efficiency and gain competitive advantage
ecosystems cover all parts of the business including policy administration systems,
rating/pricing, data solutions and connectivity to insurance partners and aggregators.
Examples of these vendors are Acturis in commercial lines and SSP, Open GI and CDL
82%
See Improved
in the personal lines space.
Efficiency

57%
Benefit from
Automated Processes
Acturis has 14 business modules that empower the
brokers to manage business. The modules include
e-trade which plugs into the commercial market to
49%
Notice Increased
seek terms of cover; and automation that reduces Competitiveness
the burden of routine policy administration such as
renewals.

41%
Record Improved
Employee Productivity

171
South & Latin America perspective
South & Latin America

How is the market driving

technology adoption?
Daniel Torres
 Majority of insurance policies are sold via agents and broker, Innovation & Business
of which a significant portion is related to motor business Development Manager
 Large SME market is still unattractive to serve via Bogota - Colombia
traditional agency models due to high product complexity
and slim margins
 The industry is starting to adopt technology to increase the
low insurance penetration despite compulsory insurance
being a feature of most markets in Latin America
SME P&C market has potential to
 Digital transformation is driven by legacy system generate significant insurance revenue
decommissioning and creating a digital distribution if the high cost of servicing this market
presence (e.g., broker portals for fully digital products)
 Insurers are starting to understand the importance and use
can be addressed with technology.
of data. Discussion on data analytics use cases are still in
their infancy

172
South & Latin America perspective
South & Latin America

Use Case Realytix Zero is an automated underwriting platform

from Munich Re. We partnered with a major insurance
The platforms self-configuration capabilities were
leveraged for simplified product development and policy
company in the Chilean market to launch a fully digital issuance through API integration into core systems.
product for the SME segment. Growth in SME product lines are being delivered for the
insurer as a result.

173
East Africa perspective
Africa

How is the market driving

technology adoption?
 This is a highly heterogenous market in terms of economic
Mthabela Sipho
models, regulation, GDP, population size, market Head of Africa Strategy
segmentation and urbanisation
Johannesburg - South Africa
 Distribution is mainly achieved via traditional channels
such as agents and broker networks together with
bancassurance
 Telecommunication industry is very sophisticated in
some markets (e.g., Kenya), which underpins many other
Successful insurers will know how
technological developments, for example enabling money to tap into the tech-savviness of
transfers and insurance to be offered in a “cloud”
ecosystem. Traditional payments such as credit cards, young people that have been pulled
and bank accounts are mainly accessed by the wealthier into the mobile money culture.
consumer groups
 Numerous market opportunities exist, including in the
DRC, due to the recent liberalisation of the insurance
industry

174
East Africa perspective
Africa

Use Case Volume of transactions (in bn

Kenya is among the leaders for fintech in Africa. The
transactions) on M-Pesa platform
telecommunication market is open for free market competition. in Africa 19,9
The local market leader Safaricom – backed by international
FY 2017–2022
shareholders – provides reliable connectivity, paving the way for
competitive offerings e.g., M-PESA which is a mobile phone-based
money transfer service, payments and micro-financing service. 15,2
Due to the popularity of their mobile platform, M-PESA leverages
12,2
their high regional mobile phone penetration to serve as a digital 11
distribution channel for insurance, adjacent financial products, and
payment transfers (similar to the Chinese WeChat platform). 8,9
Thereby, M-PESA already reaches > 50 million customers, many
6,4
previously unbanked.

2017 2018 2019 2020 2021 2022

175
South Africa perspective
Africa

How is the market driving

technology adoption?
 Complex retail products are sold by agents and brokers.
Filipe Nunes
Technology is being used to streamline onboarding and Managing Director of Life
improve risk selection via these channels
South Africa and Sub-Saharan Africa
 Mass market products continue to be very simple and
bancassurance may be able to unlock value in this segment
but legacy systems prohibit the pace of development.
Consequently, developments in mobile phone services
technology are being used to drive insurance services As the market is dominated by agent
innovation
and broker distribution, insurers are
 A key market feature is the success and maturity of loyalty
programmes used to increase customer engagement and leveraging technology to support this
build insurance brand loyalty. Successful loyalty programmes channel, and improving the engagements
need to be hyper-personalised and are dependent upon both
technology and access to relevant data with customers via loyalty programmes.

176
South Africa perspective
Africa

Use Case
One of South Africa's largest insurers launched an underwriting option in its automated
underwriting system allowing brokers to choose a streamlined underwriting journey that
bypasses medical tests for certain benefits and levels of cover.
Even though the product pricing has increased, it allows brokers to easily bundle life
insurance into product recommendations that could span life, non-life and investment
products. This facilitates a more consistent administrative burden across products for
the broker.

Where I shop 67%

The products I buy 39%
36%
How loyalty
Where I buy fuel
Loyalty programmes are
Where I bank 32%
programmes All of these 17%
providing key insights on
how to retain customers
influence The places I eat 17% and increase spend in the
behaviour How I travel 17% insurance value chain.
Where I buy insurance 6%
None of these 4%

Source: The 2022 South African Loyalty Landscape Report Nov 2022. 177
Europe perspective
Europe

How is the market driving

technology adoption?
 Most markets in Europe are still focused on traditional distribution Declan O’Neill
channels, varying amongst agents, broker and bancassurance
Head of Digital & Innovation
 Regulation, custom and practice, and maturity of European markets Strategy, Europe and Latin America
differ heavily. Accordingly, the usage and acceptance of technology
varies between countries and markets
 Smoothening and simplifying processes in the overall value chain
is key and insurers are heavily investing in this field
The fusion of technology with content rich
 There has been limited success to date for new digital native
market entrants trying to leverage data and advanced analytics propositions and data are essential for the
 In Life & Health propositions there is a clear trend towards ongoing development of new insurance
developing ecosystems which provide a multitude of tailor-made
value-added services
ecosystems.

178
Europe perspective
Europe

Use Case
Munich Re in Iberia is partnering with a digital health platform to offer insurers customised insurance products that include holistic digital health
services and outstanding customer experience. Each proposition can be fully individualised for the specific needs of the target group.

A new insurance proposition

for the silver economy

Life & Health Segmented Digital

Insurance cover Services
 Hospitalisation allowance  Unlimited video & chat
consultation (general
 Surgery covers for most
practitioner or specialist,
common surgeries from the
insurance agent)
age of 50
 Prescriptions
 Critical Illness out of country
(i.e., top treatment cover for  E-Commerce (Medical
C.I. abroad) examinations for age 50+)
 Accidents for seniors  Digital physiotherapy
 Funeral expenses  Customised e-workouts
 All content adapted for seniors
179
Next Generation
Insurance
Next Generation Insurance
Next Gen Insurance

Next Gen Insurance Study

In an online survey conducted in February 2023, we asked 97 people in Germany

Viola
between 20 and 30 years of age questions about their views and ideas regarding
the future of insurance. On the following pages, the key findings are conveyed.
Krems
Purchase
Apprenticeship Training
Integrated Studies Munich Re
insurance policy
completely
58 answered answered

digital?
(yes/no)

Submit claim
through web 82 answered answered
Tech-oriented, always online, health-conscious and application?
environmentally aware, that is how I would describe (yes/no)
the Generation Z (GenZ).
Also known as the post-Millenials, GenZ are demanding innovation in the insurance
Require personal
customer service
68 answered answered
industry. Focusing on convenience, simplicity, transparency and trustworthiness of in addition to ...
insurance products, as well as underwriting models, they insist on strong adaptability
(yes/no)
to new products and circumstances. GenZ values data privacy and sensible
processing of personal information. Fostering sustainability throughout all areas
Yes No
prevails and is key to meeting the new customer demands. A further consideration for
adapting to this new generation of customers is the importance of new advertising Percentage of study participants aged 20-30 who (1) are willing to purchase their
channels and media strategies to attract the young audience, especially through insurance policy completely digitally/online, (2) would submit their claim through an
Social Media. application, (3) require personal customer service in addition to the automated claim
handling process.
181
Next Generation Insurance
Next Gen Insurance

Perception of Insurance
To GenZ, insurance tends to be confusing, expensive and unnecessary. Products are
chosen based on familiar models and recommendations from close relatives. While
reluctantly acknowledging the need for insurance, the products promoted lack
comprehension along with the benefits for the insured. This results in the demand for
sustainable, highly customisable, service-oriented and transparent insurance products.
More concrete examples of this could include personalised, usage-based automobile
solutions and those supporting the use of sustainable vehicles. In the health insurance
sector, customisable premiums based on additional health data provided by the insured
is another option. Advancement in the service area could include on-demand, mobile-
first insurance or digital self-services. Generally the products need to be easily
comprehensible and customisable for the GenZ insurance customer.

Megatrend – Sustainability
Insurance companies need to develop a sustainable approach towards insurance
products and services to align with GenZ's values regarding environmental and
social issues.
Sustainability is not limited to the product level but rather interpreted as a development
that needs to be driven by insurers through support of sustainable practices,
renewable energy and social causes.

182
Products
Next Gen Insurance

Which insurance products is GenZ interested in? Which devices would you insure?
Do insurance companies need to adapt to the Survey results from Munich Re and ERGO Next Generation Insurance Study
younger generation or are they up to date?
This is only the starting point of questions insurers must consider. We investigated
the technological progression and labour market evolution as they offer key 73
information about the demand origin of GenZ towards the insurance market. 67
GenZ was born into a digital world which moves fast and is highly adaptable. Laptops,
smartphones or smartwatches are found in almost every household. Though the
insurance industry offers coverage for the most common devices, GenZ is one step
ahead: If I would like to insure my smart-gadget-of-the future e.g., smart refrigerators,
mirrors or wearables, is there a policy for me? The insurance industry needs to be
prepared to constantly adapt to newly developed gadgets and devices. Beyond the
question of device insurability, the protection of stored data is of immense importance.
The younger generation highly cherishes the sensitivity and personal value of the 15
information entrusted to their electronic gadgets. New insurance solutions, especially 10
7
those that address GenZ’s concerns for data protection and cyber security, will be highly
demanded in the future.
Apart from the constant emergence of new technologies, another variable to be factored Smart Home Computer/ Smartphone Smartwatch Other
is the strong shift in the labour market driven by GenZ. As remote work becomes more Devices Laptop
common, there may be a greater need for covering accidents that occur at home, as
opposed to on the commute. Similarly, as more workers become entrepreneurs, a
greater need for insurance products that protect against business-related risks such as Survey conclusion
liability, property or reputational damage, surfaces. Health insurance remains important, The findings of our study reveal that computers, laptops and smartphones
as the new generation prioritises wellbeing and mental health. are the devices GenZ participants require most to be covered by insurance
Overall, as the labour market and ownership of technologies and data changes, the role policies. Notably, as multiple answers were permitted, the most frequent
of insurance remains exceedingly important. Insurers need to challenge the selection made by participants was the combination of these three devices.
appropriateness of current product portfolios for the next generation of customers.
183
Products
Next Gen Insurance

Which insurance type is most important to you? Key Insight

Survey results from Munich Re and ERGO Next Generation Insurance Study

Growing importance
of mental health
Household insurance 5
coverage is driven
Life insurance 16
Total
by GenZ.
42

97
Third Party Liability insurance
Our survey asked
Mental Health insurance 4 respondents to rank
Car insurance mental health offerings by
insurers on a scale from
Other 8
“not important" to “very
important”. The results
show that a quarter of
Survey conclusion respondents view mental
health coverage as very
According to the results of our survey, Third Party Liability insurance was
important, 47.4% say that
identified as the most significant insurance policy by 43% of respondents,
followed by Car and Life insurance. It is noteworthy that Mental Health it is “important”, while only
insurance was regarded as of lesser importance among the surveyed 3.1% of the people asked
participants. However, this observation may not be entirely representative see mental health options
of the broader tendencies observed among GenZ. as “not important”.

184
Marketing & Communication
Next Gen Insurance

How can insurers reach Which factors are most important to you in insurance
members of GenZ? advertisement?
It is important to keep in mind that GenZ To effectively communicate with
Survey results from Munich Re and ERGO Next Generation Insurance Study.
values authenticity, transparency and GenZ about insurance, we must
social responsibility, so being honest speak their language, address their
and genuine in marketing messages is values, and prioritize digital
crucial for building trust and loyalty. For experiences that offer convenience,
instance, having a strong social media
accessibility and authenticity.
presence is crucial for reaching out to
GenZ. Any content created should consider Total Product Presentation 14
Simeon Peters

97
values and interests that resonate with
them. Furthermore, advertisement needs Junior Underwriter, Munich Re Detailed Explanation 35
to be easily understandable since the
topic of insurance can be unfamiliar and
Short and Simple 48
overwhelming for many people, especially for young, first-time buyers. Lastly, but most
importantly, insurance companies need to focus on marketing via mobile channels.
GenZ is always connected to their mobile devices and therefore mobile-friendly
websites and applications that make it easy to access, purchase and manage their
insurance policies on the go should be a priority.

In conclusion, there is no single and standard approach to reaching new GenZ insurance Survey conclusion
customers. One might take the above-mentioned forms into consideration, though the
specific demands vary strongly from person to person within the GenZ. Overall, this
The results of our study indicate that most participants prefer insurance
means staying open and attentive on all communication channels, having flexible advertisements that are short and simple. Of the remainder, 35% of
and personalised solutions for problems and developing products and services aligned respondents expressed a preference for more detailed explanations,
to customer needs. while 14% showed a preference for product presentation.

185
Marketing & Communication
Next Gen Insurance

How GenZ’s shop: The conscientious and Where are you most likely to gather information about
informed customer insurance products?
When it comes to buying insurance, the decisions of GenZ are heavily influenced by Survey results from Munich Re and ERGO Next Generation Insurance Study
technology and online resources. In general, they tend to prefer online channels,
such as insurance company websites, social media platforms, and insurance
comparison sites. An interesting difference can be seen in the participant group of our
62
survey which suggests that personal channels remain slightly more popular. This
observation supports the fact that GenZ consider themselves multi-channel shoppers, 49
and will, if necessary, embrace traditional channels such as in-person meetings or
41
phone calls. They are comfortable using various approaches to find insurance policies 38
best suited to their needs and preferences. 28
In terms of trust, our survey confirms that in addition to taking advice from family and
friends, GenZ's place high value on transparency and authenticity, and are likely to
9
refer to reviews and ratings from online sources, such as review sites and even social
media influencers. When looking into new insurance options, members of GenZ tend to 1
explore informative media in the form of short videos or social media posts that explain Family Friends Website of Insurance Social Online Other
the benefits of insurance. They also value additional online tools such as calculators to and the Insurer Consultant Media Comparison
compare coverage types and their costs. Colleagues Portals

Survey conclusion
The responses of study participants suggest that family members are the most likely
source of information about insurance, closely followed by insurance company websites
and comparison portals. It is noteworthy that the combined online channels, including
insurance company websites, comparison portals, and social media, accounted for 99
out of 228 votes. This is only slightly lower than the total number of votes received for
personal and traditional channels, such as family, friends, colleagues, and insurance
consultants, which totalled 128 votes.
186
now!
Tech Trend Radar 2023
Contributors

The Tech Trend Radar Team Supported by

Martin Thormählen Jonas Roth Daniela Gonzales Daniel Grothues

Dr. Esther Nagel Viola Krems Patrick Hinze Heiko Baumann

Maximilian Heinrich Daniela Schuler

Bill Burrows Andreas Dill

Michael Hild Kristina Tewes-Goergens

Dr. Oliver Maguhn

Hermann Schwarzmeier Norman Thoms Ingo Schenk

Alexander Wieschnewski Dr. Andreas Nawroth Johannes Benedikter Maximilian Gersbeck

Oliver Pfaffel Dr Matthias Kaper

Jakob Weiglein

Andreas Schumacher
Halyna Vasylevska
Markus Tradt

Latifa Mahna
Graham Rolland George Xu

Michal Nejthardt Dr. Andreas Bayerstadler

Marc Buchel

188
Imprint
Tech Trend Radar 2023
Contact Munich Re
Martin Thormählen
Chief Technology Officer
mthormaehlen@munichre.com

Contact ERGO
Daniel Grothues
Chief Technology Officer
daniel.grothues@itergo.com

now!
189