What are Assistant Bots

Case Study
This case study was carried out on a Healthcare provider firm. As part
of organizational strategy, the company created a shared platform to
manage process automation, innovation and better customer experience.
This was all orchestrated using BPM (Business Process Management)
and hosted a number of processes.

To understand RPA and the benefits associated with it, company started
by evaluating and prototyping this automation technology on some of its
customer business process – Capital Approval process.
Figure 1 is the AS-IS process. Once a case is approved by all roles, then finalizers logged in to the
approval digital tool to open the case and at the same time create case in JDE system with relevant detail
taken from approval tool and generate the JDE IO number. This whole process is manual and often leads
to error.

Figure 1: Capital Approval AS-IS process.

Figure 2 is the TO-BE automated process. The finalizer’s role activities were assumed by a software robot
(RPA). After the case approval is done by other process roles, the robot accesses the data store, creates the
case in the JDE system and generates JDE IO number, copies the JDE IO and pastes it on the approval tool to
complete the case and generates notification for case completion.

Figure 2: Capital Approval TO-BE process with RPA.

Results
To evaluate results of TO-BE process implementation, capital request were monitored in two sets, one set with
finalizers and another set without finalizers (Finalizers replaced with Robot) as finalizers activities were
performed by robot. The measure used for evaluating the results was case duration for fixed number for
requests. With RPA, overall duration reduced by 62% and finalizers duration reduced drastically by 99.99%.
The results may vary based on the nature of the process.
 RPA bots uncovered
RPA bots are software robots. Since they are programmable algorithms, bots in RPA
don’t have any futuristic physical form as industrial robots: they perform tasks only in a
digital environment.

The goal of bots in RPA is to automate repetitive mundane tasks and execute work
processes that are normally performed by regular human employees. Therefore, they are
often called “digital workers” or a “digital workforce.” These digital workers can work
independently as well as under human supervision (more about automation types later in
the article).

Now, let’s imagine Bob, who is a Head of Sales. The end of the month is a hot time for
Bob because he needs to calculate monthly bonuses for his team. To collect information
on deals and activities for each sales rep, he has to switch between a CRM system and
multiple Excel spreadsheets. There are 15 people in his team, so the process can take his
whole workday. However, after implementing an automation bot, it takes just a few
minutes:
 What is CRM Software? Customer
relationship management (CRM) is a
technology for managing all your
company's relationships and interactions
with customers and potential customers.

Since it’s the RPA bot that takes care of the bonus calculations, Bob now has an extra eight hours to
spend on more meaningful tasks, such as negotiations with important prospects or mentoring his team.
How do bots work and what can they do?
There are different types of processes RPA bots can handle, from simple to quite complex. Typically,
companies implement RPA for automating linear rule-based activities, such as conducting surveys, filling in
forms, invoicing, etc. Regardless of the exact function of each bot, their workflow includes four main steps:

1. Triggering
2. Input data acquisition
3. Input data transformation
4. Output delivery
1. Triggering
A trigger is an event, condition, or a set of them; simply put, it’s something that kicks off an automated task.
What will launch the bot? How is it going to run? Here are some of the most common trigger types:

• Schedule trigger. You can launch a bot on certain days or at a certain time. You can also schedule
repetitive launches, say, execute a task every three days or once a week. For example, you can
schedule a bot to prepare an executive report with analytics data at 10 a.m. every Friday.
• Email trigger. You can make a bot monitor your inbox for emails that meet certain criteria, for
example, if there are any attachments. Once the requirement is met, the bot kicks in.
• File trigger. A bot launches when there are certain files or folders available, for example, if new files
are added or their amount reaches a certain number.
• Window trigger. This launches a bot when a certain window opens or closes. It can be useful to set a
trigger for an error message pop-up window.
• Interface trigger. A click, a keystroke, or a sequence of keys trigger a task. You can also make this
trigger work only in a certain app.
• Database trigger. The task kicks off when a specific operation happens in a database, for example, if
someone enters new data.
2. Input data acquisition
Before acting, a bot needs to collect certain data. It could be information from emails,
spreadsheets, CRM and ERP systems, databases, etc. As a rule, modern RPA platforms
enable OCR, or Optical Character Recognition. This technology recognizes text and, for
example, can retrieve it from relevant fields in scanned documents to use for further
processing.
Optical character recognition (OCR) is the extraction of data from scanned files and PDFs and its
transformation into machine-encoded text. OCR is an area of research in artificial intelligence, pattern
recognition, and machine vision.

The sources of data might vary from scanned documents, photos of a report, to subtitle text superimposed
on an image, and PDF files. For instance, passport documents, invoices, bank statements, digital receipts,
business cards, mail, or any other type of printed documentation.

OCR is extensively used as a form of data entry from printed data records. This technology allows people
to digitize data for increased productivity.

How OCR can be used

The combination of OCR + RPA allows you to redirect employees from routine tasks to analytical
assignments and to earn more money through company growth rather than increasing the number of
employees.
Using OCR, ElectroNeek can recognize and read information from accounts, invoices, contracts, tables
and other financial documents and can automatically enter it into the corporate systems of your enterprise
– Salesforce or others. It can also prepare financial transactions, build or update excel tables, send
messages to Outlook, and perform almost any other office routine.
3. Input data transformation
This step is the actual purpose of RPA bots. For example, it can be sorting emails, transcribing voice
messages, creating an invoice by filling in the fields with data taken from a variety of other applications
like CRM, and more. You’ll find plenty of ideas for automation in ElectroNeek’s Bot Library.

4. Output delivery
Once a bot executes its target action, it needs to communicate the result with stakeholders (“Here’s your
task. It’s done.”). The report can be in the form of:

- An email;
- A report in PDF, XLS, or any other format;
- A generated dashboard;
- An update in a database;
- A notification in a messenger;
- and more.
Real-life example
Let’s get back to Bob and his sales bonus calculation bot. The workflow can be easily broken into those
same four steps.

• Triggering: A person manually launches the process by double-clicking the desktop icon.
• Input data acquisition: The bot collects the latest data on activities and transactions.
• Input data transformation: The bot inserts the collected data into a spreadsheet with bonus calculation
formulas.
• Output delivery: The bot distributes the final bonus sums via employees’ emails.
What can’t RPA bots do?
The key distinguishing feature of RPA bots lies in their decision-making process. Their actions
are based on a pre-programmed scenario, which may consist of a single condition or a set of
them. That means they rely on formal logic and can’t improvise or effectively operate in
changing conditions.

Another thing to keep in mind is that RPA bots don’t strive to achieve any particular high-level
goal. They execute the processes that are explicitly specified in their task. The task is considered
completed if the bot succeeds in executing the necessary steps. Even if there are plenty of error
messages, as long as they don’t prevent the execution of the specific process, the bot will ignore
them.

The important conclusion here is that a formally successful execution of a script doesn’t
guarantee the expected business result.
Types of bots in RPA
Electroneek image
There are two basic types of RPA bots: attended and unattended ones.

Recently, we’ve witnessed the appearance of a new, more intelligent one. What is the purpose of these
types, and which one is better? Let's find out.
Attended RPA bots
Attended automation is like a virtual assistant that helps you complete tasks. The emphasis is on “helps,”
because you still need to handle the bot; now, it’s just easier for you. A good example would be a software
bot in a call center that sales reps can use to look up information while they’re talking to clients. In other
words, an employee can launch an attended bot to take over just a part of an activity. Once the bot has
finished with its part, the employee gets control back.

Attended RPA is often referred to as a “desktop robot” because it usually sits right on your computer.

Use cases for attended RPA:

• Accelerating and streamlining employee’s daily routines

• Repetitive tasks that require real-time human-system interaction
For example, Call Centers can leverage RPA to automatically answer calls and get information about
that customer from various business applications: CRM, SAP,…

Nowadays, Attended RPA will help businesses handle many different blocks of tasks: business
processes or facilitate decision making and boost productivity. However, this technology also has some
disadvantages that companies need to be aware of.

One of the most significant limitations of Attended is that it needs to be on employees’ servers.
Therefore, it is only capable of executing on a fixed server, and employees cannot simultaneously use
the software at the same time.

In addition, attending RPA also has another disadvantage that prevents the computer from being used
while it is running. Then, kernels will not perform any other tasks while the RPA is running. This
limitation can cause employees to wait for hours during RPA’s performance.
Unattended RPA bots
If you hear something about RPA bots that are ready to work 24/7, that's exactly what unattended
automation is about. In unattended automation, workflows don’t require a person to manually kick it
off: a bot is self-triggered when it meets certain circumstances. For example, it can send alerts when
your office printer requires maintenance or the ink needs refilling.

Unattended bots run on servers.

Use cases for unattended RPA:

• Automating back-office processes

• Automating tasks (and even job roles) that can be entirely mapped
One of the best examples of an unattended software robot system is an automated invoice processing system. If a
business finance department needs to automate the processing of received invoices, they can leverage Unattended RPA
to do the job.

The most significant advantage of Unattended robotic process automation is that it won’t prevent employees from
working on their computers. They can work concurrently with RPA bots without any interruption. For tasks like invoice
processing and data entry that take a long time, choosing an RPA system is a good solution.

However, these duties may involve employee confirmation requirements as unattended robotic process automation will
work automatically until it completes the task. Employees will not need to be involved in the process.
Hybrid RPA Systems/Bots
Hybrid RPA Systems is a combination of Attended RPA and Unattended RPA. Therefore, it
possesses most of the outstanding features of both types.

Specifically, you can choose to use Hybrid RPA Systems as one of the other two types. It can work
flexibly and efficiently on the same or a large server. In addition, businesses will not need to use
employees to conduct monitoring and process installation for Hybrid RPA Systems. This bot system
is capable of automating most tasks without human intervention.
What are the benefits of bots in RPA?
The benefits of integrating robots into the human workforce are numerous. They can significantly
boost enterprise productivity because they are several times faster than the average employee when it
comes to repetitive, rules-based, admin tasks. Plus, they never need a rest or lunchbreak, working
tirelessly 24 hours a day, seven days a week.

Automation of routine processes eliminates the risk of human error and significantly improves
SLAs, resulting in happier customers. Employees will be happier to, because they will spend less of
their working day on soul-crushing and repetitive tasks.

A robotic workforce can be scaled up rapidly as a business grows – there is no need to recruit or
train people. What’s more, with improved performance, better resource utilization and less time
spent fixing errors, enterprises can realize a rapid return on their investment.
Intelligent process automation (IPA)
Intelligent process automation is the most recent type of RPA. It employs artificial intelligence (AI)
technology, enabling bots to learn, grow smarter, and get better at processing non-linear tasks and
unstructured data.

A couple of examples of IPA would be image recognition or voicemail transcription, which leverages
natural language processing. Another example is handling invoices. With AI, an RPA bot can not only
process payment requests, but also structure the invoices, which may come in various forms from
different contractors, according to a single set of rules.