TALKDESK REPORT

2024 Talkdesk
global contact
center KPI
benchmarking
report
Foreword
Long before the introduction of AI, advances While some organizations may be reticent to generative AI? Is your contact center retrieving the
in technology had already sparked a shift in the adopt AI due to trust gaps and uncertain guardrails right answers? Are you asking the right questions?
traditional contact center model. Rather than a or controls, early generative AI adopters report Use this report as a starting point.
pure cost center, contact center agents established higher agent productivity, improved deflection
themselves as a bridge between a brand and its to self-service, and increased customer scores. This report benchmarks handpicked contact center
customers, transforming it into a profit center. The demand for AI only continues to climb. performance metrics applied to a wide range of
This tie has grown stronger as more brand industries, company sizes, and locations to inform
interactions take place entirely between consumers With this demand, there will be new ways critical business decisions, helping you use data
and agents. The addition of more self-service of measuring the metrics of AI in the future, as a tool for transformation. We’ve also included
options and channels gives customers increased but for now, benchmarking traditional contact examples of organizations that are seeing success
access to brands, and agents must be able to handle center KPIs and operational improvements remains today, demonstrating the impact of leveraging AI
all interactions, regardless of the channel in which highly valuable and requested by contact centers and automation in the contact center.
customers are engaging. worldwide. Following the introduction of AI solutions,
we’re seeing contact centers use these traditional
NEVILLE LETZERICH
Following years of AI innovation from KPIs as early indicators for improvements in agent
Chief Marketing Officer
forward-thinking CX vendors like Talkdesk, the performance, customer drop-off rates, and more.
Talkdesk
introduction of generative AI to the CX industry didn’t
just accelerate the demand—it changed everything— Are decision-makers more likely to embrace the value
including how customers can interact with brands. of data as we are able to make better use of it with

2
Methodology
Handpicked metrics for operation leaders
The data presented in this report is proprietary to Talkdesk.
It represents the average performance of a sample size
of nearly 3,000 clients from a wide range of industries,

3,000
across six continents: Africa, Asia, Europe, North America,
Oceania, and South America.

Clients
This data was collected from the Talkdesk platform between
January 1, 2022, and December 31, 2023.

6
Continents

3
Table of contents
I. Introduction 05

The importance of benchmarking 06

The benefits of benchmarking 07

II. Key findings and metrics 09

2023 Overview 10

III. Metrics deep dive 13

IV. Metrics by: 24

Industry 24

Region 26

Company size 27

Conclusion 28
I. Introduction
As companies are focused on creating value from With a radically changing landscape in 2023, how can
AI innovations, seeking to automate processes, contact centers identify the most effective strategies
This report analyzes the call data of nearly
empower their agents and illuminate opportunities to achieve their goals? Navigating this change starts
3,000 Talkdesk customers in different
for improvement within their data, contact centers with learning more about peer performance in 2023
locations and industries, from SMBs
will continue to track new and existing KPIs to and comparing against it.
to enterprise organizations.
measure impact.
The 2024 Talkdesk global contact center KPI
As we look back at contact center performance benchmarking report measures and examines
over 2023, including six months of data since the impact felt by contact centers in metrics such
the introduction of generative AI, we have seen as average talk time and service level. It analyzes
a significant acceleration of self-service and the call data of a sample size of nearly 3,000
automation technology, and the empowering nature Talkdesk customers in different locations and
of AI to optimize the work of the contact center industries, from small and midsize businesses
agent. This shift has fundamentally reshaped how (SMBs) to multinational enterprises. This resulting
organizations are interacting with customers, report contains a first look at these metrics,
though many are just beginning to measure providing a snapshot of the impact of the last
the impact across traditional KPIs. year on contact centers.

5
The importance of benchmarking

Contact center benchmarking is a structured, systematic process

of continually identifying, examining, deploying, and reviewing
Benchmarking allows center managers to determine best
best practices to gain and maintain competitive advantage.
practices, critically evaluate their business, and implement
As AI accelerates the transformation of the customer experience,
changes when necessary.
benchmarking ensures alignment with industry strandards
for improving CX strategies.

“I think oftentimes you have the system in place but you don’t
necessarily have the bandwidth or the capabilities to really measure.
And so leaning into the experts... you’ve invested in the software
solution, but really leaning on that team to look at the numbers from
an outsider’s point of view and just kind of make recommendations,
kind of talk you through, and want to help you improve. I think that’s
really valuable.”

— DIRECTOR, E-COMMERCE & DIGITAL MARKETING - ONLINE CLOTHING RETAILER

6
The benefits of benchmarking 2. Uncovers areas for 3. Supports testing of
improved efficiency improvement initiatives
1. Identifies urgent areas
of improvement Comparing your operational performance Contact center benchmarking allows you
to industry measures allows you to understand to assess the effectiveness of AI initiatives
if your organization falls above, below, or average, for different channels. This can strengthen
Benchmarking can identify ineffective practices
and if you are more, less, or equally affected the contact center’s competitive advantage
to help establish priorities by highlighting areas
by them. by improving practices and fostering innovation.
in which you are underperforming compared
to peers.
When assessing contact center efficiency, it can By evaluating successful— and unsuccessful
be helpful to compare the following metrics: —practices in your industry, you can leverage
When focusing on contact center effectiveness,
• Average speed of answer (ASA) the experiences of others to accelerate growth
it can be helpful to compare some of the following
• Average talk time (ATT) and adaptability in your organization.
metrics:
• Average after-call work time (ACW)
• Customer satisfaction (CSAT)
• Average abandonment rate (AAR)
• First call resolution (FCR)
• Average time to abandon
• Average revenue per contact (including
• Average waiting time (AWT) It’s also important to consider measuring
percentage of upsell or cross-sell opportunities)
• Average handle time (AHT) impact beyond contact center KPIs to
• Volume of inbound and outbound calls other performance metrics across the
business. A value calculator can help
your business assess the value of AI-
powered CX across customer experience,
workforce effectiveness, operational
efficiency, and business growth.

7
The statistics for 2023 are correlated • Total inbound calls increased
21% to 390M in 2023 across
• 52% agent attrition in 2023,
with 76% of agents saying they
to the adoption of AI solutions, new Talkdesk customers, following are overwhelmed by systems and
a similar trend in the Deloitte information. Companies currently
support channels, and the effect they Digital report where 27% of using generative AI are 35% less
have had on contact center operations. contact centers say channel likely to report agents overwhelmed
expansion has increased total by the information in front of them
Over the last year, we have seen: service inquiries. The average during calls.
speed of answer decreased
significantly, down 13% to • Average hold time increased
an average of 8.7 seconds. 7% in 2023 to an average
of 16.9 seconds.
• Average talk time increased
slightly at 4.3% for an average Over 2023, these KPI benchmarks
of 2m 50 secs in 2023, though have offered an informative insight
decreasing dramatically from into the performance, pitfalls,
3m 35 secs in 2020. and opportunities for success
in contact centers.

8
II. Key findings and metrics
In 2023, contact centers that used AI-powered insights and data
were able to remain responsive and deliver effortless customer
journeys to outperform their competitors. Contact centers are
adapting to the rapid innovation created by generative AI, KPI
for which there isn’t yet a definitive playbook. Leaders are
looking at their peers and business partners to effectively
deploy AI solutions that automate more customer interactions,
empower agents, and illuminate insights.

As mentioned earlier, benchmarking is a natural first step.

Excelling in core contact center competencies implies successfully
implementing AI solutions, resulting in a high-functioning contact
center with streamlined interaction handling, improved agent
efficiency, reduced attrition, and optimized resource allocation.

You should:
• Understand relevant KPIs for your business goals.
• Compare performance against your competitors.
• Establish benchmarks with the best ways to improve your KPIs.

9
2023 overview

81%
According to a CCW Market Study on the state of generative
AI, 81% of contact center leaders expect it to make a meaningful
impact in their contact center. 86% have plans for generative AI
investments, and more than 50% planning for meaningful action
this year.
of contact center leaders expect
These positive outlooks are already backed by data. generative AI to make a meaningful
Despite a 21% increase in call volume from 2022 to 2023, impact in their contact center.
speed to answer improved significantly and service level
improved slightly.

86%
“There is absolutely nothing more important than
the customer journey, and Talkdesk Workforce
Management streamlined our scheduling and
planning, allowing us to reduce hold times and
abandon rates when volumes spike even further
of contact center leaders
than ever before.”
have plans for generative
— KEN COHEN, SVP, SALES & BUSINESS DEVELOPMENT, JK MOVING AI investments.

10
While contact center leaders begin to see the major benefits
of generative AI, 78.6% of agents still fear AI might eliminate
“Our team leverages Talkdesk Copilot, an AI-powered assistant that
their jobs.
proactively delivers contextual recommendations, next-best actions,
and automated assistance to the agent during live interactions.
2023 benchmark data reports show that the role of the agent
It also automatically creates a summary of the call using generative
is still mission-critical. With inbound call volumes reaching over
AI, streamlining after-call work. An agent might otherwise spend
390M and talk times hovering around 2m 50 secs, they’ll likely
eight minutes summarizing the conversation themselves, so saving
focus on handling more complex issues and customers who
that time on each call delivers impactful operational gains.”
are skeptical about AI’s ability to resolve their queries.
— THOMAS GROSSO, EXECUTIVE DIRECTOR, SERVICE DESK, CAI

Even as call volumes increase, copilot (agent assist) tools

are effectively empowering agents to resolve problems faster,
with overall talk times dropping 20.5% since 2020 (originally
3m 35 secs down to 2m 50 secs).

This comes at a critical time as agent attrition remains

record-high at 52% and leaders report that an overwhelming
amount of channels, information, and systems significantly
contributes to negative agent experiences. Copilot solutions and
tools like automatic summary ease the burden of administrative
tasks for agents, all in one place, to streamline the agent
experience, from pick up to wrap-up.

11
 Metrics

Metrics 2023 2022 % Var YoY 2024 so far

January to May

Average
abandonment rate 9.3% 9.17% +1.41% 9.22%

Average talk time 2m 50s 2m 43s +4.5% 2m 57s

Average hold time 16.9s 15.8s +6.9% 17.6s

Average speed
of answer 8.7s 13.1s -33.5% 8.2s

Service level 75.61% 75.39% +0.38% 76.4%

12
III. Metrics deep dive
Average abandonment rate

# ABANDONED CALLS IN QUEUE

Due to the significant increase in calls
and agent attrition rates in 2023,
9.17%
ABANDONMENT
contact centers had to work hard to
RATE
TOTAL # INBOUND CALLS - # ABANDONED CALLS NOT IN QUEUE keep up with call volume. Callers were
willing to spend an increased average Average abandonment
of 3m 24sec waiting for an agent rate in 2022.
(a notable decrease from 2020
Talkdesk defines abandonment rate average waiting times of 4m 22 sec),
as the percentage of calls terminated and the average abandonment rate
Data point: For all Talkdesk remained flat from 2022 (9.17%
by callers after being added to a
customers in 2023, the to 9.3%). It is worth noting that while
queue post IVR (minus calls dropped
within a short abandoned call
threshold). This KPI gives an insight
into how many calls were terminated
average abandonment
rate was 9.3%.
agent attrition was significantly higher
in 2023, self-service solutions like
autopilots can help contain some
9.3%
before reaching an agent. portion of these large volumes
of calls despite smaller workforces.

Average abandonment
rate in 2023.

13
 To optimize average
Customer spotlight abandonment rate:

Collins Community Credit Union

cut abandonment rate in half—from Deflect calls with self-service options for
30% to 14.5%—after deploying the high-frequency questions and requests.
Talkdesk AI-powered CX platform.
Talkdesk Autopilot answers frequently
asked questions, freeing agents’ time Employ workforce management tools
and allowing them to help customers and improve call peak forecasting.
with more complex issues.

Automate repetitive tasks so agents

can focus on higher-value interactions.

Other notable Talkdesk customer highlights

Use in-queue time to deliver
important information.

LegalShield reduced British Columbia Lottery Apple Federal Credit Union

abandoned call rate Corporation (BCLC) reduced abandonment
to 3.5%. reduced abandonment rate from 26.1% to 14.1%
rate to 12%. in one year.

14
Average talk time

In 2023, the value of human interactions is still paramount.

TOTAL TALK TIME
Average talk time (ATT) at Talkdesk customer contact
AVERAGE TALK
centers increased slightly from 2022 to 2023, up 4.3%,
TIME
# HANDLED INBOUND CALLS though significantly dropping 21% from 2020 (3m 35 secs).

OR

TOTAL HANDLING TIME - TOTAL HOLD TIME - TOTAL WRAP-UP TIME 2022 2023
AVERAGE TALK
TIME
# OF CALLS HANDLED 00:02:43 00:02:50
hrs min sec hrs min sec

Talkdesk defines average talk time

(ATT) as the amount of time an agent

4.3%
Data point: For Talkdesk
spends talking with customers,
customers, the average
excluding other activities an agent
talk time was 2:50 minutes Slight increase of
might need to complete an interaction.
in 2023. average talk time (ATT)
at Talkdesk.

15
The dramatic decrease from 2020 may be attributed to the To optimize average talk time:
increased adoption of copilot tools that help agents solve complex
issues faster, while a significant portion of simpler calls are now
being handled via autopilot tools across digital and voice. Drive proactive training based on areas
highlighted from AI-powered insights.
On its own, ATT doesn’t provide a complete picture of the quality
of service. For example, when an agent spends more time speaking
with a customer, the negative impact on ATT is often outweighed
Leverage faster responses with intent
by increased customer satisfaction (CSAT), if that time is spent
detection and generative AI call routing.
fully resolving an issue. If the end goal is to improve customer
experience, the negative hit to ATT is a worthwhile investment.

Simplify agent effort with copilot tools

to assist agents.

Key takeaway: Having a higher or lower average

talk time is not inherently good nor bad as AI and
automation tools get more complex issues to agents;
its importance lies in the impact on average handle time
(AHT) and customer-facing metrics (CSAT, NPS, etc.)

16
Average hold time

TOTAL TIME INBOUND CALLS WERE ON HOLD

AVERAGE
HOLD TIME
# TOTAL INBOUND ANSWERED CALLS

Talkdesk defines average hold time as the measure Putting or leaving callers on hold hurts the customer
of time customers were put on hold when on a call experience. Given the higher call volumes in 2023,
with an agent. Hold time shouldn’t be confused with average hold time increased alongside other metrics
wait time, which is defined as the time a customer in 2023, rising to an average of 16.9 seconds, only 1.1
has to wait before an agent answers the call. seconds higher than in 2022 (15.8 seconds). Longer
average hold times correspond to an increase in call Average hold
length (when compared to 2022 data). time (AHT) in 2023.

Data point: The average hold time for

Talkdesk customers was 16.9 seconds,
in 2023.
16.9
seconds

17
This increase in average hold time may be the Together, average hold time and average talk
result of high agent attrition in 2023, and the time impact average handle time (AHT). A high
overall increase in call volume from 2022 to 2023. average hold time negatively affects AHT and
As stated earlier, many contact centers have overall contact center efficiency. Common causes
deployed self-service digital support to handle for leaving a customer on hold are searching
simple, frontline issues, which frees agents to for additional information or transferring a call to
resolve more complex queries. a more suitable agent or department. Implementing
copilot solutions automatically gives agents answers
to customer queries. Optimized call routing can
reduces average hold time by getting customers
Key takeaway: Decreasing AHT will to the right places quickly, positively impacting
reduce contact centers’ cost per contact the customer experience.
driven by the ability to handle more
interactions with the same number
of agents. To drive down AHT without
hurting the customer experience, first
analyze the current AHT and the total
wrap-up time.

18
 Customer spotlight
To optimize average hold time:
Pogust Goodhead reduced wait times
and boosted their capacity to interact Give agents AI-powered tools that
with customers, moving through automatically deliver the right answers
outgoing calls quickly and processing in real time.
incoming calls effectively, resulting in
savings of $750K across the business.
Leverage faster response with intent
Instead of taking several minutes to detection call routing.
confirm a caller’s identity and locate
the appropriate claim, Talkdesk CX
Cloud delivers calls to the right agents Identify the root causes of long hold
in seconds. An innovative CX platform times by leveraging AI to illuminate insights
empowers their teams to work more hidden across large volumes of call data.
effectively and efficiently.

Help agents and customers help

themselves with an AI-powered
knowledge base.
Other notable Talkdesk customer highlights

Apple Federal Credit Union reduced British Columbia Lottery Corporation

average hold time from 146 to 103 (BCLC) lowered average hold time
seconds. to 24.7 seconds.
19
Average speed of answer

TOTAL TIME CALLS SPENT RINGING

AVERAGE SPEED
OF ANSWER
# ANSWERED CALLS (IN SECONDS)

Talkdesk defines average speed of answer (ASA) Even though contact centers experienced
as the amount of time it takes agents to answer significantly higher call volumes in 2023, the average
an inbound call, including the time when the agent’s
phone rings but excluding the time the caller spends
in IVR or queuing.
speed of answer improved significantly from 2022,
with the average call getting picked up by an agent
in just 8.7 seconds (a 33.5% decrease from 13.1
8.7
sec
seconds in 2022). This reduction may be an indicator
that more customers have deployed automation
to engage with customers faster. We expect this
Data point: On average, the average number to decrease with further adoption of AI
speed of answer for Talkdesk customers and intent detection.
was 8.7 seconds, in 2023.

20
When contact centers improve ASA, there is less time between To optimize the average speed
the phone ringing and the agent answering the call. While this
improves agent productivity, it has less impact on CSAT since the
of answer and average wait time:
caller’s journey started when the number was dialed and routed
through the system and queued until an agent was available. Deploy self-service autopilot across
channels to reduce unnecessary calls
and drive more productive conversations.

Decrease agent attrition by examining

Key takeaway: Rather than looking at the average
common causes for turnover, including
speed of answer as a stand-alone metric, tie it together
evaluating the systems and tools that can
with the average wait time to determine the total time
contribute to a negative agent experience.
a caller spends before reaching an agent.

Evaluate your IVR strategy and consider

how generative AI can connect customers
to the right place, faster.

Evaluate different customer requirements

on waiting time and diversion strategies.

21
Service level

# ANSWERED CALLS + # MISSED INBOUND CALLS WITHIN THE THRESHOLD

SERVICE
LEVEL
# ANSWERED CALLS + #TOTAL MISSED INBOUND CALLS
75.6% 78.12%
Talkdesk defines service level (SL) as the percentage Service level is the holy grail of contact center
of calls answered and missed within a predefined performance and customer service quality
threshold, e.g., 20 seconds. SL measures the management. Overall average service level remained Average service level Service level in business
availability of a company to its customers. flat from 2022 to 2023— landing at 75.6% in 2023. in 2023. services in 2023.

Notably, some industries exceeded this metric

including business services at 78.12%, RCG at 78.5%,
and manufacturing at 80.45%. Industry-tailored
Data point: The average service level for solutions, like Talkdesk Autopilots for Retail, Banking,
Talkdesk customers was 75.6% in 2023. and Healthcare, are pre-packaged to accelerate
AI adoption and can help organizations exceed
benchmark service levels. 78.5% 80.45%

Service level in RCG Service level in

in 2023. manufacturing in 2023.

22
 Other notable Talkdesk customer highlights
Customer spotlight
Talkdesk customer Tuft&Needle achieved
a service level of 96% of calls answered
within 20 seconds. With a service level Employbridge reduced LegalShield increased
below 90% at the start of their journey, human-assisted contacts by service level by 4%.
the Tuft&Needle team sought to improve 20% with a 90% service level.
their contact center performance with
better technology and clearer visibility
into performance.

Through the insights available through

Talkdesk, the team discovered that agents Key takeaway: Service level can provide a high-level
were available at the wrong times, leading overview of contact center performance. AI-powered
to missed interactions. With adjustments insights, customer metrics, and other KPIs provide more
to team coverage, Tuft&Needle surpassed detail into where agent productivity and efficiency can
their service level goals. be improved.

23
IV. Metrics by
Metrics by industry

Average
Average talk time Average hold time Average speed
Industry abandonment Service level (%)
(minutes) (seconds) of answer (seconds)
rate (%)

Agriculture 6.84% 3m 18s 3.1s 22.5s 71.47%

Business Services 9.52% 2m 21s 10.4s 8.6s 78.12%
Consumer Services 4.49% 2m 40s 26.3s 5.3s 74.8%
Education 16.63% 1m 43s 2.7s 8.9s 74.17%
Financial Services
& Insurance 9.06% 3m 4s 23.9s 7.5s 75.01%
Government
& Public Sector 9.79% 2m 43s 6.3s 7.5s 70.74%

24
Metrics by industry

Average
Average talk time Average hold time Average speed
Industry abandonment Service level (%)
(minutes) (seconds) of answer (seconds)
rate (%)

Healthcare
& Life Sciences 6.74% 3m 26s 19.6s 8.9s 75.68%
Industrial Manufacturing 6.15% 3m 55s 20.7s 11.5s 80.45%
Media, Communications
& Internet 11.08% 2m 4s 12.9s 7.1s 71.35%
Others 12.89% 3m 29s 14.2s 10.7s 64.29%
Retail, Ecommerce
& Consumer Goods 8.18% 3m 19s 18.5s 7.6s 78.45%
Telecommunications 13.7% 2m 6s 22.4s 7.8s 63.26%
Travel, Transporation,
Hospitality & Logistics 11.39% 2m 24s 18.5s 14.7s 76.07%

25
Metrics by region

Average
Average talk time Average hold time Average speed
Location region abandonment Service level (%)
(minutes) (seconds) of answer (seconds)
rate (%)

Africa 12.55% 52.5s 12s 6.7s 59.03%

APAC 12.42% 1m 51s 10.6s 4.4s 76.94%
EMEA 10.17% 2m 35s 15.1s 9.6s 75.32%
East Europe 9.97% 3m 12s 8.8s 10.7s 76.63%
West Europe 10.18% 2m 33s 15.3s 9.6s 75.29%
LATAM 6.74% 2m 45s 23.1s 4.4s 75.67%
NAM 8.75% 3m 12s 18.2s 8.9s 76.57%

26
Metrics by company size

Average
Company size Average talk time Average hold time Average speed
abandonment Service level (%)
(# employees) (minutes) (seconds) of answer (seconds)
rate (%)

1-5 12.61% 3m 9s 15.7s 8.8s 72.75%

6-25 10.54% 2m 12s 8.8s 9.5s 77.03%
26-50 7.79% 3m 2s 12.4s 7.8s 82.68%
51-100 8.67% 2m 41s 11.1s 8.7s 77.86%
101-200 12.21% 2m 40s 13.8s 9.6s 70.26%
201-500 8.1% 3m 9s 20.1s 8.8s 76.5%
501-750 8.75% 2m 53s 15.7s 9.2s 75.96%
751-1000 9.1% 3m 22s 23s 7.7s 76.3%
1001-2500 10.4% 2m 27s 15.3s 7.5s 73.8%
2501-5000 10.5% 2m 21s 13.5s 8.3s 75.4%
5001+ 7.62% 3m 25s 25.4s 9.5s 76.69%
27
Conclusion
We have seen that 2023 was a year like no claims – companies need to partner with a vendor
other; generative AI has transformed customer that offers practical applications of AI with industry
expectations, and smart enterprises are leveraging expertise, specification, and proven results. A close
its power to match those expectations with AI-fueled partnership and proven, innovative technologies
customer experiences. While organizations have allow companies to meet customers’ ever-evolving
different structures, strategies, and objectives, KPI expectations to deliver continuous, high-quality
benchmarks provide an informative view into contact customer experiences.
center performance, reveal trends, and uncover
opportunities for improvement. Going forward, Discovering how your organization ranks among
we must leverage the collective experience peers is crucial to the decision-making process.
and learnings from implementing innovative AI Use this business intelligence as the competitive
solutions. With the increase in AI and automation advantage that drives your business to reach
solutions in the market – and equally as many AI its goals on the journey towards transformation.

28
 Drive exceptional business outcomes with AI-powered
customer experience technology built for your industry.

Talkdesk® is on a mission to rid the world of bad customer experience. With our cloud-native, generative AI-powered CX platform,
purpose-built industry solutions, and extensible AI offerings, we empower enterprises in the cloud and on premises to deliver
exceptional customer experiences that make them more competitive, grow revenue, reduce costs, and provide operational
efficiencies. Learn more and take a self-guided demo at www.talkdesk.com.
+1 (888) 743-3044
www.talkdesk.com © 2024 Talkdesk, Inc. All Rights Reserved Worldwide. Unauthorized use, duplication, or modification of this document in whole or in part without the written consent of Talkdesk, Inc. is strictly prohibited.