RPA-Robotic Process Automation

Key learnings>>>

Software robots will increase speed, agility, and efficiency. It will replace human workforce with
digital software bots who will perform high-volume, repeatable, rule-based monotonous task

There are 2 types of software robots

1) Unattended Autonomous- Does work without human intervention

2) Attended/Assistive- Used by Human operator for tasks which require high level thinking.

Uses of RPA-

1. Perform frequent, repetitive, consistent data entry, or execution of well-defined series of step.

2. Inputting & or synchronizing data b/w multiple systems "Swivel Chair"

3. Supervise data entry work of human against set of rules.

4. Software robots used for data migration, extracting & reporting.

5. Software robots can be used for rule-based decision making. For e.g., going through complex
data to reduce it down to key decision.

Benefits of RPA to business-

1. Software robots will increase accuracy, consistency & speed.

2. Software robots can handle very repetitive, low valued task inexpensively.

3. Instantly scale up/down to meet demand.

4. Work will be getting done 24/7

5. No holidays, mistake or "bad days"

6. Full digital audit trail of tasks & steps. With that trail we can drive performance & business
metrices even if it wasn't originally built into the system

Benefits of RPA for employees

1. Enable human to work better, smarter & more creatively on higher value tasks

2. Increases job satisfaction because of not doing monotonous work.

Example of application of RPA in business processes

1. Finance department-

Accounting, Budgeting & Payroll etc activities which might involve copying data from one place to
another & making reports can be done by software bot.

2. Marketing Department-

Software bots can deal with customer service, automatically respond to new leads that are
generated & automatically create new advertisements based on the stimulus in outside market.

3. Human Resource Department-

Software can perform rule-based processes pertaining to recruitment, attendance & staff training.

Most widely tool used RPA- UiPath RPA tool

What are key drivers for RPA

1. Need for constant company growth (Reduced cost+ Increase in revenue= More profit). Any cost
saving directly goes to bottom-line.

2. Labor arbitrage has been exhausted

Three types of labor arbitrage

a) Worker proactively relocates to other country

b) Employer sponsor immigration via providing temporary visa

c) Outsourcing/Offshoring job to foreign country

Challenges like strict immigration policies, increase in wages in low wage countries due to tech &
infra development in those countries. Most low wage countries have been used up.

But software bots can perform work cheaper or at par with outsourcing with a smaller number of
human resources.

3. BPO challenges

a) Potential border tax on services

b) Immigration reform

c) Reform to H-1B visa process

d) Time zone delay which affect Productivity

e) Less communication, alignment & maybe more number of FTEs

Work will move onshore & managed by Human interfacing with digital workforce.

4. IT changes can be prohibitive

a) IT is mostly busy working on new, core competence business ideas.

b) IT system integration are risky & could create new problems.

c) IT changes are expensive, require many layers of people & organization.

d) IT changes take too long due to availability, complexity, testing etc.

e) IT staff might lack skill & knowledge required to modify legacy system.

Hence business can implement RPA using existing systems, with no to minimal system integration
work. There is overall less involvement of IT.

5. More work as companies has easier access to customer via internet & smartphone.

6. Increasing customer's expectation (better, cheaper & faster)

7. Competitive global marketplace require agility & speed.

8. Employee job satisfaction & retention (as business are competing for same Human resources)

9. Better RPA tools are available in the market.

Will RPA kill outsourcing?

Ans is maybe, it boils down to performing a cost- benefit analysis. Even BPO's are implementing
RPA to remain relevant & become more efficient.

Oversea bot might cost equally to local bot.

So outsourcing is surely going to change:

a) Now it becomes a conversation about value.

b) Contracts maybe renegotiated to “Outcome-based"

c) BPO would be paid %age of savings created by automation. Meaning less use of FTEs with
increased productivity. It is different to existing where BPO tries to maximize FTEs to improve
revenue.

Right mindset for RPA automation:

Any effort to improve efficiency / cut cost should not reduce Quality

One should strive for excellence

Robot Benefits

1) Reduced cost

2) Improved speed

3) Improve Quality

4) Improved consistency

Robot v/s Human

Computers are faster, accurate, reliable for repetitive rule-based task

Software robot can assist & replace human workforce/operator

Employee can now focus on high value task

Those employees will be retained by the org who are high on knowledge on technology & have
automation skill so they can create & update automated task.

Best fit for RPA

Not restricted to any specific functions/industries

There is guardrail for type of process that yield best Return on investment.

• Repetitive
• Rule-based
• Standardized
• High Volume

Qualifier for RPA

• Desire to handle more task in more time-efficient manner

• Desire to grow while keeping expenses low
• Copies of information in multiple systems
• Legacy system that are hard, expensive &/or it is risky to change
• Regulated

RPA can be a key differentiator for small business too.

Best fit industries for RPA are:

1. Insurance- It has a lot of time-consuming high volume repetitive task.

RPA can free up 20-30% of Human capacity.
It can reduce policy processing time by 40-80%
 2. Banking & Financial- They have significant numbers of transactions & data volume
They are very regulated industry: so, auditability, security & accuracy is crucial & those
things Robots are better than human
Banking has old legacy system which are outdated & risky to update. RPA can help banks to
speed up without risk & high cost of complicated IT integration.

3. Legal- Highly dependent on record keeping. So, they have lots of paper documents.
RPA can use OCR-Optical character recognition to scan, catalogue & search/correlate
documents faster than human.

4. BPO-Business Process Outsourcing-make own company efficient & add automation value
to the organization it serves can benefit BPO industry.

Benefits, Challenges & Risk of RPA.

Benefits-

• Potential for rapid ROI

• Improve accuracy, speed & security
• Reduce human capital costs
• Faster scaling
• A 24/7 virtual workforce
• Better exception handling
• New analytics & insights
• Job Satisfaction

Challenges-

• Choosing the right processes to automate

• Setting realistic expectations
• Employee resistance
• ROI calculations become more complex
o Robot license costs
o Process optimization & requirement gathering
o Salaries & time: Development, testing, monitoring & maintenance
o License splitting
• Deployment
o Software installation
o Permissions/entitlements
o Training already busy people
• Software update logistics -software vendor who produces RPA software/tool will make changes to
improve their tool, so org will have to be plan ahead for upgrading their
o Orchestration server
o IDE (Development environment which the developer use)
 o Robot agents – new versions of actual robots’ agents both on intended robot server
& human agents’ machine.
• Robot maintenance-
o Monitoring/Timeliness-Robots won’t always work perfectly so we need to monitor
them & if failure does occur then we need to tackle in timely manner.
o Intuitive logging (screenshots)- Line by line traces of steps they are performing &
taking screenshots & storing them on server along with link to those SS within log.
So If something is goes wrong then troubleshooter can quickly click on link & view
the SS to get a better sense on what had gone wrong// failed.
o Process knowledge- People who have Process knowledge if question arise what if
robot move to unknown situation. Such people can really help in such a situation.
• Script Organization-
o Centralized vs. Decentralized- If a business team listen to what vendor say then
different business team might be using different tools & have different scripts
spread across the organization. This informal development of robots can lead to
confusion & inefficiency of their own.
o Version control- Involve IT to do version control. Version control is IT best practice
where script or programming files are stored on a central server which protects
them from being accidentally deleted if an individual computer crashes. It also
allows multiple people to contribute to the script file without having to pass around
multiple copies of the files.
o Reusable workflows- Consider using a RPA architect with processes SMEs to know
when certain workflows can be reuseable from robot to robot. Those workflows will
be broken into separate different files & version control them in a way that makes it
easy for other robots to use them but you do not break those robots if someone
decides to update one of those centralized assets.
• Change management-Thought on change management to ensure robots evolve along with a
changing process/system. We need to have a way to notify developers that the system/process is
changing, so that robot developers can modify the robot at the same time when the system &
process is getting change. So, we do not wind up pushing a new process/system into production &
have it break the robot.
• Finding the right balance of IT involvement & formality- Many vendors suggest that they get these
RPA tools & start automating their processes, it is true to some extent, but it is going to be very
important & sometimes difficult to find the right balance of IT involvement & formality while building
a robot. Many business teams are frustrated with the slow pace of IT & a sense that it likes to
overcomplicate things. But a good enterprise, RPA endeavor is absolutely going to be partnership
between business & IT resources.
• Handling unstructured &/or handwritten data- RPA have problem with handling unstructured & or
handwritten data. The latter is most problematic because even the best OCR engines aren’t
necessarily great at reading different kinds of handwriting. Unstructured data means things like
emails & text documents that aren’t structured in a way that a database table or an excel
spreadsheet is. Vendors are working hard at producing artificial intelligence that soon be able to
make sense of unstructured data like that. So, it could be challenging to read such data hence it is
very important to have correct process selection before going ahead with RPA. Make sure for the
 process you are planning to implement RPA should not have things like unstructured & handwritten
data.
• Daytime savings time- Sometime daytime savings time can throw a wrench into your RPA program.
Unattended robots are typically scheduled in some kind of an orchestration server & if these
unattended robots need to run at specific times during the day, you may find that those times
during the day will change with daylight savings time. But the scheduling server itself does not
factor that in & suddenly your unattended robots are running an hour too soon or an hour too late.
So it is important to put an process around daylight savings time & having somebody make sure that
the schedules are good during those times of the year.
• Password management- Many companies have passwords that expire within 30/60/90 days &
sometimes you can work around that with internal passwords, but in many cases you will be
dealing with external systems as well over which you will have no control. So, it’s important to have
some kind of governance & process where you have got somebody responsible for changing those
passwords in a timely manner. This can be in-built into robots. Robot can detect a condition &
accordingly change its password.
• Compliance- We have to think about it- can a robot replace or supplant humans for regulated
processes. There are many regulated processes that follow stringent government requirements &
rules. It might require an organization’s legal team to evaluate whether a robot can replace human
being in those kinds of circumstances.
• Complexity- Like unstructured & handwritten data discussed previously. Many manual tasks have
eluded automation for a long time due to the steps involved which machines won’t be able to deal
with. But it is expected that with time RPA tools will eventually be able to do complex tasks & such
tasks will become fewer & fewer.
• Human Resource- After automation we have to think about things like which people to keep. What
are their roles going to be & how will they be measured? As we know, with automation we will need
fewer people that too with different skills which are currently present.

Risks-

• Errors- Possible to program the robots wrong

• Security Risk- Danger of Putting clear text credentials into their scripts while implementing
automation. They might think it will only run on their machines but at some time the robot
might be emitting log files or logs into server location somewhere & suddenly you find those
credentials are being passed around within the company. Involving IT can help to mitigate
security risk because IT people deal with issues like security on a regular basis.
• System overload- Robots can generate more volumes than humans, if we deploy robots
suddenly & they are doing work 24/7 then there will be chance that vol of work they generate
can overwhelm downstream systems.
• Redundancy- Lack of redundancy is a risk. Whom to approach if a robot fails to do a
specific task as the humans who used to perform that task earlier is gone. Mitigation for this
is to create excellent documentation around the process.
• BPO “IP lock in”- BPO will use RPA as a selling point to try to seduce companies who aren’t
necessarily willing to do their own RPA. Problem will be for those BPO’s who have developed
their own proprietary RPA tools, as it can lock an organization with a BPO simply because
now you are depending upon their proprietary tool to execute your process. Whereas if one
 had developed the robot themselves using industry standard tools or use BPO that also use
industry standard tools. This will give more freedom to choose any other BPO to work with.

Automation Maturity Levels

FUNCTION KNOWLEDGE & RELEVANT TECH
RULES
Stage 1 • GUI & backend • Statis rules set • Web & email
Acting Basic interaction interaction
Digitization • Quality & Testing • Screen
• Desktop Scrapping &
automation OCR
• Document
workflow
• Task scheduling
Stage 2 • Basic decisions • Static rules set • Enterprise
Enhanced • Analyze • Contextual search &
Digitization unstructured knowledge insights
data • Natural
• Named entity language
extraction processing
• Sentence • Content
segmentation analytics
• Speech tagging • Process
automation
Thinking Stage 3 • Understand • Dynamic rules • Cognitive
Cognitive customer capability computing
decisions • Query rules • Extensive • Data mining
engines knowledge • Pattern
• Complete • Needs Real-time recognition
transaction data feed

Current RPA tools are primarily focused on Stage 1 & 2. Tool vendors are working hard to
develop new capabilities to facilitate Stage 3.

Enterprise RPA Goals

Any enterprise starting an RPA endeavor should keep 3 things in mind.
1. Focus on Stage 1 & stage 2 wins
• Stage 1 for quickest wins
• If you don’t have in-house technical capability, the there are service providers
offer PoC (proof of concept) before commitment
2. Develop an automation strategy
 • RPA offers rapid, noninvasive deployment
• Should not translate to tactical, opportunistic problem solving
• Develop an automation strategy that aligns business with IT
• Attack it as you would any other strategic program
3. Don’t ignore post-deployment reality
• What will it look like when you have 1000 robots executing transactions?
• How will you monitor performance & failures?
• Factor in TCO- Total cost of ownership (including training) & Maintenance

Intelligent Automation journey:

Acting
Capability Outcome
Robotic process Automate application using GUI & static rules
automation

Data Science & Derive insights about performance & expand predictive
Analytics analytical capabilities

Cognitive Automate judgement work on unstructured data &

Computing develop machine learning

Thinking & Acting Digital workforce Adapt analytics & strategic insights to drive enterprise
flexibility & value

Current/Best RPA Automation tools

High level tips to keep in mind while looking for an RPA tool.

• Try to find a RPA tool that presents a holistic solution which means it offers an IDE-
Integrated development environment, ideally it should offer scheduling & reporting
server, it would offer intuitive logging capabilities & bots themselves that can be
installed on human agent machines or back office servers.
• The tool one chooses should be ubiquitous, that simply means that it has a wide user
base with both commercial and user support options. User support- online forum
where many users who use the tool are frequently posting & answering questions for
other users. Ideally use tool that is used & supported by major implementation
partners, companies like Deloitte, PricewaterhouseCoopers etc.
• Ease of use is very important – Make sure tool has clean, intuitive user interface. Ideally
one should be able to drag & drop activities onto the workflow.
Easily arrange and rearrange & configure them.
 It should offer a recorder which makes it easier & faster for you to capture the flow of
the process, & then go back & optimize it with variables & looping & so on.
Offer parallel scheduling & execution of robots
• The solution should be flexible, meaning it should
Support automation across many targets (web, MS Office(word/excel), automate API
calls, DB(database queries),ingest documents via OCR, perform screen scraping & data
scraping, control mainframe terminal, send & check emails)
Bonus if solution integrates with cloud services & there are multiple cloud services that
offer OCR or language translation or sending text messages via SMS & so on.
Tools should also allow one to create custom-coded activities using language like VB.NET
or C sharp, so that you can work with IT to develop an activity that does something
complex that tool does not offer right out of the box. While can help less technical use to
just drag & drop that custom coded activity into a workflow & easily configure it using
the property pane.
• Solution should be scalable & utilize a robust underlying framework. A perfect example
of that is a tool like UiPath uses the Microsoft workflow foundation framework, which is
a very stable & well-known mechanism for automating business processes that is
produced by Microsoft.
• Make sure the solution is secure & it has various encryption features, especially for
encrypting things like website credentials, application login credentials & so on.

There are companies who evaluate all the tools which are available in the market & make
recommendations to organizations. Everest group is one of such companies. According to them
leader in RPA tools are: Uipath, Blue prism & automation anywhere are top 3.
Another organization Forrester Group have also mentioned above organizations as the leader
under the category.

RPA roles & Job-

Enterprise RPA roles- Roles fit under 3 different categories-

• Leadership
• Robot development
• Robot Operations

Leadership Robot development Robot Operations

Sponsor RPA COE Leader(s) Virtual workforce supervisor
Champion Business Analyst Robot Operator
 Change Manager Architect Robot service & support
Steering Committee Robot Developer Infrastructure Engineer
Robot Tester
Infrastructure Engineer

Leadership-
1. Sponsor-
Typically going to be a senior executive, might be from IT, Finance or operations. Ideally
this person would be the one to initiate the RPA program. But RPA can also initiate from
a grassroot initiative & work its way up. Ideally RPA should be well understood & driven
from top-down.
The sponsor will support & protect the adoption of automation throughout the entire
business.
Regardless of their “Day-job” this person must be passionate about RPA’s potential for
improving cost, accuracy & compliance on a modest timeline & budget compared to
typical solution deployments. IT integration takes too long and costs too much money &
introduce too much risk.

2. Champion
Could be an IT or operational leader. They can be head of COE operations.
This person will be responsible for the health of the automation pipeline.
They will evangelize benefits & maximize popular support among users. It is important
for troops on the ground to really see the value of automation & not be fearful of their
jobs.
This person will set & manage expectations as to what software can accomplish.

3. Change Manager
It is often someone from HR. Ideally, they have an executive presence, which means they
are capable of influencing people throughout the organization. They have excellent
situational awareness, so they can kind of get a read on how employees are generally
feeling about RPA & adjust their strategies accordingly.
They should be able to balance conversations & build consensus.
This person will ideally drive adoption & acceptance of technology-based change
initiatives.
They will determine & deploy the training necessary to achieve the “future-state” model
which is some mix between human workers & digital workforce.
They will define communication plans & facilitate deployment of them.

4. Steering committee
 A dedicated & well-rounded team of individuals that can understand challenges,
opportunities & benefits for the entire organization & not just specific departments.
All these diverse team members have a common picture in their minds & they will
ensure that the scope, direction & outcomes of RPA projects are in line with your org
needs & expectations.

Robot Development-
• Business Lead
o Business Analyst
• IT Lead
o Solution Architect
o Robot Developer
o Robot Tester
o Infrastructure Engineer
o
1. Business Analyst-
o Process SME & highly engaged with your target business users
o They will identify, quantify & document opportunities for automation
o Ideally they can highlight compliance nuances that may impact automation
choices
o They might also do UAT user acceptance testing once a robot has been
completed

2. Solution Architect-
o It’s a technical role & one of the primary things they will be responsible for
is determining on a sliding scale what percentage of a process should be
done with IT integration versus RPA(automation bandaid), v/s traditional
manual setups. It might turn out that some processes are best suited for a
100% IT integration. Some processes might benefit from 100% RPA & some
processes might be a mix b/w 2 or more of these things
o Once determination is made they will define the solution architecture, & if
they are conversant with one or more than one RPA tool, they might even
choose which RPA tool is most appropriate for the particular process being
considered.

3. Robot Developer-
 o It’s a pretty straightforward role. They will design, develop & possibly test
robots according to the solution architect’s vision

4. Robot Tester-
o Test robots during the development processes, but it is optional because
the robot developer might be sufficient for testing.

5. Infrastructure Engineer-
o It is going to be critical to handle tool installation & upgrades, backend
integration, which means things like getting permissions for databases &
APIs & they might also be responsible for managing virtual worker
entitlement which means creating user accounts on virtual machines that
give the virtual worker the ability to access the system being automated.

Robot Operations-
1. Virtual Workforce Supervisor
Will be responsible to scheduling , monitoring & fine tuning your virtual
workforce
o Unattended Robots
o Attended Robots
We are primarily talking about unintended robots here, but the attended robots
might also emit logging events & generate metrics from which we can derive
insights that would allow a supervisor to make decisions & maybe schedule things
differently or scale them differently on multiple machines things like that.
2. Robot Operator
Pretty obvious role & these are existing business users who are triggering
unattended robots to accomplish some percentage of the manual tasks they used
to do.

3. Robot service & support

It will be frontline support for deployed RPA solutions.
They should have strong triage, RPA & troubleshooting skills & these people
might even qualify to be RPA developers, but they are more focused on fixing
existing robots instead of building new robots
One might need multiple people on call if we have a digital workforce that’s
performing 24/7-7 day a week
 4. Infrastructure Engineer
They are both on the development side & on the operation side. It is a crucial
role. They are dealing with tool installation & upgrades, making sure backend
integrations are still operational & of course assigning & updating virtual worker
entitlement where necessary
It can be an on- call rotation position

RPA JOBS -also fall under #3 categories

• RPA software vendors

o Software Engineer
o Pre-Sales Engineer
o Sales- Non-technical position but still the person has to be conversant
about RPA to be able to suggest a product
o Support- who can handle a constant stream of questions that will
inevitably result from the BPOs & consuming companies using the tools.
Software engg & presales engg will have higher technical skill than the developer
found in service providers or RPA consuming companies.
• Service Providers/BPO
o Pre-Sales Engineer- They know how to do RPA well & can help BPO
salespeople with RFPs & handling proofs of concept & generally doing the
heavy technical lifting to remove any roadblock that might prevent signing
a deal to provide RPA services.
o Practice Lead-Once BPO does sign a deal & provide RPA services for a client
, they will need a practice lead to handle the implementation for one or
more clients.
o Solution Architect
o RPA Developer
There is a good chance that the BPOs might also have business analyst & process
SMEs but that’s probably up for negotiation because the company they are
helping out might have their own process SMEs & business analyst that will work
with the service provider to get the job done
• RPA consuming companies
o Architect
o Developer- the one who needs to know the most about using RPA tool.
o Traditional project delivery roles (Business Analyst, Infrastructure engineer,
Support etc.)
RPA Job sites
• Indeed.com
• Monster.com
• Dice.com
• Linkedin.com
Search terms to look for
• RPA
• Robotic process automation
• UIPath
• Blue Prism
• Automation Anywhere

Section4: Implementing RPA in your enterprise

Overview of implementation Automation:

Enterprise RPA
• Determine the processes to automate
• Quick discussion about attended versus unattended robots
• Running an RPA proof of concept to get the program off the ground
• Creating automation COE to help mature & scale your RPA effort
• Provide tips about change management that might help to smooth out some
bumps along the way
• Summary: tips for success

Criteria for Ranking RPA ideas

Which process automation will give best ROI is a great thing to do. Choosing right
processes is fairly straightforward

• RPA isn’t restricted to any specific type of process

• There are “guardrails” regarding processes that automate best
 • They should be repetitive & predictable, which means they use the same buttons,
& fields on the same systems over & over again
• They should be rule based, which means they do not require any human
interpretation
• They should have structured data, which means the data should come from
something well formatted, like excel spreadsheets or database tables or even API
calls or tables on a website
• The number of transactions should be mid to high volume, RPA can be valuable if
the volume has peaks & troughs, because dealing with that task manually means
that you would have to hire enough people to deal with the peaks, & keep them
sitting around idle through the troughs if you didn’t want to have to release them
& hire them back for peak cycles
• You get the most bank for your RPA buck if your processes can be handled during
normal operating hours, & outside of normal operating hours. Sometimes IT
organizations will shut down certain systems to do technical housekeeping or
backups or whatnot, & that may prevent your software robot from being able to
work on that process after say 7:00 pm. So that maybe something that can be
negotiated with the IT team or it may not. So it is great to consider when your’re
choosing processes to automate.
• You should evaluate whether a particular process is prone to human error
• Also think about whether the process involves duplicate data on multiple systems
because that’s an RPA sweet spot.
• The process itself should not be the subject of ongoing process optimizations, &
the system that the process uses ideally will not be going through changes either
• Be sure to factor in the amount of time it will take to actually develop the robot
because the faster you can get robots into production, the faster you’re gonna
see return on investment

Selection Process:
Basic selection process you can use for choosing automation candidates
• Determine which criteria you’re going to measure for each process.
• Develop a stack ranking excel file that has those criteria as columns
o You can optionally add additional columns that will provide a weighting
factor to increase or decrease the weight of a particular criteria.
 • Gather up the business stakeholders & maybe even the people who perform the
process.
• Ask those people/stakeholder to evaluate the criteria for each process on your list
• Finally once you have rated all the criteria for all the processes then you can sum
them up as scores & use the EXCEL RANK formula to see which process ranks the
highest & that you’ll subsequently automate first.

RPA pocess selection

Off-hours processing

Quick developmemt
Structured Outputs

Volume variability
Structured Inputs

Process stability

System Stability
Duplicate data
Total volume
Rules-Based

Error Prone

Score

Rank
Process A 1 1 1 2 2 5 1 5 2 2 2 24 2
Process B 1 2 1 1 4 1 3 4 1 1 1 20 5
Process C 2 1 3 4 3 1 2 3 1 4 1 25 1
Process D 3 2 2 5 1 1 5 2 1 1 1 24 2
Process E 1 4 1 4 1 1 2 1 5 3 0 23 4
Ranking Scale= 1 to 5
Above is excel stack ranking spreadsheet that you could help to determine which
process to automate first,
Rows – list of processes A to E.
Columns- Several criteria, you can put other columns too: e.g., Potential of this
process to save time, or maybe number of FTEs that would be displaced by
automating this process, or even the potential for increased revenue by
automating this process.
As we interview the stakeholders who actually person these processes, so we
would simply ask them to use ranking scale 1-5 but we can also use 1-3 or 1-10
whatever we felt was most applicable.

ROI-Building the business case-

Additional criteria & concept that will help to refine the selection list
• When you first consider introducing RPA into your enterprise many people are
going to be hyper-focused on the cost avoidance aspects of this technology. That
simply means spending less money on human resources.
 • Successful business cases should be more complex than just labor reduction.
• What we mean by above is that R in ROI should be more than just money. Have
conversation with senior business leaders & find out what other criteria are also
important beyond cost avoidance, things like-
o Quality of work, reducing the error rates
o Speed at which tasks get done, potentially freeing up humans to do other
things throughout their day
o Agility the ability for businesses to pivot & react to different ideas &
market pressures
o Insight The ability to understand your business better
o Retention if your workers are doing monotonus task all days they are not
going to be happy at their job
o Compliance- Maybe the fact that you are offshoring work has created
compliance issues due to new regulations & the data you have been
sending offshore need to come back onshore.
o Higher throughput- It can mean an increase in revenue for the business
• Even after you have that conversation, your senior business leaders are most
assuredly still going to wanna know about that money.
Things to be kept in mind when you’re considering automating any process.
Current process & cost
Let’s say we have got a process in our business that we are considering automating, & at
present six human workers are being paid $600 a month.
The cost of this process is $3,600 a month i.e., $ 43,200 a year. Let’s say for sake of
argument, each of these humans accomplish 100 tasks per day & error rate is 15%
Software robot Options
It is important to realize there are several possible paths you can take to automate a
process depending upon the nature of the process & what’s important to the business. If
the only thing important to the business is cost avoidance, that’s going to influence your
choices. But if there are other considerations such as increase throughput & greater
compliance, you may choose a different path.
Some options for the process we’re currently looking at are:
1. Unattended robots Onshore- We can use one unattended robot Onshore, which
would mean we have a computer or VM onsite that might cost us $250 a month
 for license fees. This scenario would get rid of 6 human workers or possibly have
them upskilled or retrained to do some other task
2. 1 Unattended robot+ 1 Agent Offshore- This could be another option, where
unattended robot does the majority of the processing of these tasks & maintain 1
human agent around who handles exceptions that the robot just can’t do. This
decision will be influenced by how stable & consistent the process is. If you just
can automate it 100% with an unattended robot, you may have no choice but to
keep a human around to handle the exceptions
3. 2 Agents with attended robots Onshore- Human agents are equipped with
attended software robots. We have chosen to move these onshore because
maybe there was a major compliance issue with our data being offshore. And
additionally, maybe doing this task during local work hrs instead of offshore work
hrs will make it better for our customers. So this choice is being driven by what is
most important to your business stakeholders. In this scenario, there’s a chance
that the throughput may decrease because we are only using 2 agents instead of
6. The fact that they’re using attended robots makes them significantly more
productive than normal humans, but the impact on your throughput remain to be
seen.
4. 6 Agents with attended robots offshore- In this final scenario, if business revenue
is dependent upon the volume of these tasks being completed, maybe the senior
leaders decide they want to keep the six agents but give them attended robots to
improve their performance. In this case, we won’t be spending any less money,
but you also would not be spending more money in hiring additional humans to
scale up your output.
So, having that conversation with your senior leaders to understand what’s truly
important other than cost avoidance is going to help you to make decisions about
the RPA approach you’ll take.

Let us say we had a conversation with senior leaders for this process, & they have
told us that they want to reduce the costs, improve the timeliness for which these
tasks are completed, meaning they prefer that it happens in local time. They’ve
also indicated that there are new compliance needs, which means they would
prefer having the data onsite instead of offsite. And of course they want to reduce
the error rate & rework of transactions.
So, for those reasons we decide we are going to eliminate the 6 offshore humans,
use 2 human onshore which doubles the cost of each resource but allow us to get
the work done in a more timely fashion on local hrs. And we are going to equip
these 2 human agents with attended software robots to increase their output &
reduce the error rates & maybe some aspects of this process require human
intervention, which forced us to use few human workers instead of just using
unattended robots.
Our cost is now $ 2,600 per month, which is the 2 humans plus the licenses for
unattended robots & that equals to $ 31,200 a year.
So out cost avoidance is $ 1,000 a month or $ 12,000 a year which isn’t too bad.
As it turns out, these 2 humans equipped with software robots are now capable
of producing 300 units a piece instead of 100 before, & error rate has reduced to
2%. So we’re still getting the same throughput with few errors.

So we have achieved cost avoidance & maintained our throughput, decreased

error rates, got the data onsite & tasks are being performed in a more timely
fashion on local hrs.

So as the planning was occurring for producing these robots, it’s easy to think we
can simply compare the cost of 6 agents at $20 an hour each against the cost of 2
agents at $40 an hour each plus the licensing cost of the robots. And like any
business case we will project it out for 3 years.
So in 3 years, 6 human agents would could us 129,600 whereas 2 onshore agents
which are more expensive using attended software robots would only cost us
93,600.
Our total saving over 3 years= 36,000 (12,000 per year)
Before
Item 2018 2019 2020 Total
Total
6 Agents($20/hr.) 43200 43200 43200 129600 cost

After
Item 2018 2019 2020 Total
2 Total
Agents($40/hr.)+Robots 31200 31200 31200 93600 cost

So above might not be a incredible cost avoidance number, we have to remember

that we’re getting the benefits of:
• Lower error rate
• Keeping our data local
• Faster response that happens on local time
 • Improving retention of employees – as software robots make the work less
drudges.
Above was just an example of a business case, actual case can look as below:
Before
Item 2018 2019 2020 Total
6 Agents($20/hr.) $ 43,200.00 $ 43,200.00 $ 43,200.00 $ 129,600.00
SW, HW, License etc
($100/month) $ 7,200.00 $ 7,200.00 $ 7,200.00 $ 21,600.00
$ 151,200.00

After
Item 2018 2019 2020 Total
2 Agents($40/hr.)+Robots $ 31,200.00 $ 31,200.00 $ 31,200.00 $ 93,600.00
SW, HW, License etc
($100/month) $ 2,400.00 $ 2,400.00 $ 2,400.00 $ 7,200.00
Requirement analysis (40
hr, @$60) $ 2,400.00 $ - $ - $ 2,400.00
Robot development (80 hr,
@$100) $ 8,000.00 0 0 $ 8,000.00
Robot testing(4 hr/month,
@$50) $ 2,400.00 $ 2,400.00 $ 2,400.00 $ 7,200.00
Robot maintenance(4
hr/month, @$100) $ 4,800.00 $ 4,800.00 $ 4,800.00 $ 14,400.00
Redocumentation (20 hr,
@$50) $ 1,000.00 $ - $ - $ 1,000.00
Trainer of agents (10hr.
@$60) $ 600.00 $ - $ - $ 600.00
$ 134,400.00
So this bring our 3 year cost avoidance= $ 16,800 ($ 5,500 per year)
So clearly our cost avoidance is lower than we might like, but we must keep in mind that
we are still getting those benefits we talked about-

• Timely execution of tasks

• Local data
• Decreased error rate
• Increased employee retention

Let us check a process which allows us to 100% automate the process with no human
intervention
So, it will allow us to replace 6 human resources with 1 Unattended robot as we
discussed in previous cases. This robot can be a VM or physical computer in our onshore
office. It’s realistic to think that the license fee for this unattended robot could be $250 a
month.
Its important to remember that a software robot can work at least 2 times faster & 3
times longer than a human being. (2 times faster figure is quite conservative, it’s
actually 3 to 5 times faster).
So, in the above scenario our cost is $ 250 a month, i.e., $ 3000 a year. Hence our cost
avoidance is $ 3,350 a month i.e., $ 40,200 a year. Since robot is working 2 times faster &
3 times longer than humans, that one unintended robot is still capable of producing 600
units per day, & error rate is now down to 0%.

Before
Item 2018 2019 2020 Total
6 Agents($20/hr.) $ 43,200 $ 43,200 $ 43,200 $ 129,600

After
Item 2018 2019 2020 Total
1 unintended Robot $ 3,000 $ 3,000 $ 3,000 $ 9,000

The total savings over 3 years would be $ 120,600. Apart from cost avoidance we also
get the following benefits:
• Quality with 0% error rate
• Compliance by keeping local data
• Improved the speed i.e., faster response time. It may be possible a single robot
might work faster than 6 people.
• If we built it correctly then we might be able to add a second unattended robot &
increase our throughput & revenue even more without incurring a lot of
expenses.
True cost
Before
Item 2018 2019 2020 Total
6 Agents($20/hr.) $ 43,200.00 $ 43,200.00 $ 43,200.00 $ 129,600.00
SW, HW, License etc
($100/month) $ 7,200.00 $ 7,200.00 $ 7,200.00 $ 21,600.00
$ 151,200.00

After
Item 2018 2019 2020 Total
1 unattended Robot $ 3,000.00 $ 3,000.00 $ 3,000.00 $ 9,000.00
SW, HW, License etc
($75/month) $ 900.00 $ 900.00 $ 900.00 $ 2,700.00
Requirement analysis (40 hr,
@$60) $ 2,400.00 $ - $ - $ 2,400.00
Robot development (80 hr,
@$100) $ 8,000.00 $ - $ - $ 8,000.00
Robot testing(4 hr/month,
@$50) $ 2,400.00 $ 2,400.00 $ 2,400.00 $ 7,200.00
Robot maintenance(4
hr/month, @$100) $ 4,800.00 $ 4,800.00 $ 4,800.00 $ 14,400.00
Redocumentation (20 hr,
@$50) $ 1,000.00 $ - $ - $ 1,000.00
Trainer of agents (10h,
@$60) $ - $ - $ - $ -
$ 44,700.00

Our cost avoidance for 3 years is still over $ 100,000 which translates to $ 35,500 a year.
And here too we have other benefits as well.

Summary:

• Begin by asking the business what’s important in addition to money. Those could
be factors like
o Quality
o Speed
o Agility
o Insight- Understand your business better
o Retention
o Compliance
 o Higher throughput (increase revenue)
• Let those desire help determine which RPA approach to take
o If they only care about money, you wouldn’t choose to Onshore
o If response time & local data are important.. onshore is better.
• Do not overlook robot development & operating expenses in ROI calculations
(Manual & robots)
• Measure metrics for the “before” process
• Build the robot to emit telemetry that covers the “after” metrices (qty & speed)
• Create a dashboard so the business can see quantified improvement that has
resulted from RPA program

Qualifiers Benefits
Off-hours processing

Improve compliance

Improve Throughput
Quick development
Structured Outputs

Improve Retention
Volume variability
Structured Inputs

Process stability

System Stability

Improve Quality
Cost Avoidance

Improve insight
Improve Speed

Improve Agility
Duplicate data
Total volume
Rules-Based

Error Prone

Score

Rank
Process A 1 1 1 2 2 5 1 5 2 2 2 2 5 5 0 4 5 2 2 49 1
Process B 1 2 1 1 4 1 3 4 4 1 1 1 4 4 0 4 3 4 1 44 4
Process C 2 1 3 4 3 1 2 3 4 4 1 1 3 3 0 3 3 1 3 45 3
Process D 3 2 2 5 1 1 5 2 4 1 1 5 2 1 4 1 5 2 1 48 2
Process E 1 4 1 4 1 1 2 1 4 3 0 4 1 1 5 1 2 1 4 41 5
Ranking scale = 1-5
Through great conversations, you can probably use your gut instincts to put ratings on all
these criteria & wind up with the list of processes that’s most likely to get you the
business outcome you desire.

UiPath automation Hub business case

As we have used excel spreadsheets for our automation ideas, it would be better if we
use better tool to keep track of our automation ideas. So one of those is Uipath
Automation Hub
Uipath Automation Hub is by far the most comprehensive & powerful way to manage
RPA ideas. It has so many capabilities
• One can add more information to their ideas once they have added it on Path
automation hub.
• Automation hub also involve automation score which one get by answering the
questions in the high level assessment.
• There is also a detailed assessment section where more criteria are present.
• We have a cost benefit analysis tab, automation hub generates a Y1 & Y2
reflection of the estimated cost savings comparing to original manual cost. It also
gives you place to track your project plan for analysis, solution design,
Development, Testing & so on.
• It further gives you a documentation tab for each idea that gives you
downloadable templates or the PDD (Process Definition Document), Solution
Design Document, Development specification document & so on. It allows you to
add files including the templates as discussed as well as things like detailed work
instructions, process maps & even the automation source code that was
developed to satisfy this idea.
• Collaborators tab is also there, so you can see which member on the team were
responsible for submitting the idea, who’s business analyst, who is the RPA
developer & so on.
• If we go to edit mode then we can assign the idea to be developed by CoE or by a
citizen developer power user, or even a citizen developer self user. We can keep
track of which phase any given idea is in. From assessment, to review, to Analysis,
solution design, development & testing. We can keep track of the status as well.
• Above & beyond the capabilities for each idea, we also have a people tab where
you can view a list of members in your automation program. You can see the
components that were developed & uploaded that can be reused on different
automation ideas. And they even incorporate gamification that provides badges &
achievements & ranking, which can potentially keep people more engaged in your
automation program, resulting in recognition, awards, and so on.

Creating RPA proof of concept- Guidelines

• Consider creating 1 attended & 1 unattended robots (to experience the difference
between the 2 & have a sense for what they entail.
• Choose 1 or 2 relatively simple automation processes candidates.
• Decide whether to do RPA work yourself or make use of RPA vendors
o One pro in favor of vendors is that they’ll often have tools & licenses ready
to go. That can save some upfront time when your’re just getting started &
keep your POC short.
o Another pro is that the vendor shouldn’t need to come up to speed on the
tools. They should already have people on staff who know the tools well. If
your internal team member need to learn the tool that can certainly take
some time & make your POC longer than it needs to be.
o One potential con of using a vendor is that they might provide biased
advice to increase your dependency on them, or maybe even cause you to
spend more than you would have if you just did everything yourself.
• When you are doing POC , you’ll want to choose 1 to 3 tools so you can compare
how easy they are to use & how effective they were at automating your process.
You might be surprised that one tool would be significantly easier & more
intuitive to use than the others.
• Decide whether to try it on one computer or involve your IT team to get VM and a
server and so on. Better to use later approach because it’s always great to get IT
involved early & everybody can have a better sense for what you’re trying to
accomplish & what the process will look like in the future when you try to scale.
• Once you have got everything you need & somebody on the POC knows the
process you’re going to automate really well, go ahead and create the robot &
take notes along the way about what it’s going to take to scale it in the future.
What it means that maybe when you build robots in the future, you’ll need to
have multiple environments & create test data so you don’t have to test using
production data. Make note of how easy or hard it is to purchase & apply
licenses, things like that.
• Once the robot is completed, make sure you test the robot or robots you’ve built
on a variety of data samples. Oftentimes we think there’s very low variability in
our process, but when you ultimately try it on a robot that operates on very rigid
rules, you find that certain data samples might cause the robot to throw an
exception. So at that point refine your robot’s logic so that it has capabilities to
deal with wider variety of data.
• After testing is complete , go ahead & evaluate the ROI from multiple
perspectives like speed, ease of use, reliability & how repeatable this process is
going to be in the future.
• Produce succinct, high quality narrated videos to demonstrate the value of what
was produced. This is something you can distribute to key stakeholders so they
 can get a much more clear visual about what software robotics look like & you
can even deliver it to team members to get everybody more excited about the
potential for RPA.
• Once POC is completed & everybody agrees that there’s value to proceed with
RPA, then it’s time to formalize & create an automation center of excellence

Create Automation CoE

Bringing together the right people, process, & tools to ensure the most successful
enterprise-grade RPA.
Starting small- One person can download an RPA tool & perform an RPA POC. Then
individual business owner could certainly follow suit & just start automating any
relevant processes.
Risk taking above approach-
• Possible zero sum game
• Chance robots will block manual work because they are running on business
people’s local machines
• Chance to wind up with duplication of common scripts.
• More expensive tool licenses.
• Inconsistent scripting approaches
• Disjointed metrics
• Security issues
• Lost work
• Decreased likelihood of success
Center of excellence- Establish it & bring together the appropriate people, process,
&technology , there’s much better chance that your RPA endeavor will succeed.
 People/Structure Process
•Right people with •An effective &
right skills in the repeatable
right role process to deliver
high quality
applications

Technology
•A perspective set of
enterprise best practices &
policies

A center of excellence is simply an organizational construct to ensure that __RPA___ is

successful & efficient.
A center of excellence is a single point of accountability.

Why we need CoE i.e., Center of Excellence

• COE will establish a framework for getting work done.
• It will bring a sense of governance, which means
o Coordinating & controlling action & resources
o Oversight which means preventing people from doing the wrong things.
o It covers de-duplication & synchronization
o It can help to resolve conflicts when they do arise.
• CoE can be a centralized source for Infrastructure- They can install & upgrade
VMs, RPA tools, & servers.
• COE can be responsible for centralizing knowledge & information, maybe in
systems like sharepoint, confluence & so on.
• COE will establish scripting conventions & reuse best practices.
• COE will help to assess security, meaning which system can various bots access.
Establishing RPA means creating a virtual workforce, & each of the robots should
be treated similarly to human resources when it comes down to permissions.
 • CoE will develop, test, run, monitor and fix robots. Running here refers to
unattended robots. Business team will be the people running the attended
robots.
• CoE will help establish exception processes. Like what do we do when the robots
don’t do what we expect?
CoE Tasks
• Publish documentation of marketing , onboarding & the various processes that
are required for getting work done
• They will establish & execute a procedure for ranking & selecting process
candidates
o Determining what %age of manual v/s RPA v/s IT integration will be most
sensible for any given process
o Help to uncover synergies b/w processes & efforts. This is where a
duplication may arise, if everybody was doing their own thing
• CoE will help to pick the RPA tools.
• They we’ll also install & configure the RPA tools & the orchestration server, &
they’ll also be responsible for the disaster recovery infrastructure in case there’s a
catastrophic failure with the main infrastructure.
• They will likely choose & configure the demand management tool. It could be
something like Jira or Rally or Botstatz. Ideally Robot development should be a
agile process, so it’s great to choose a tool that supports the agile workflow.
• They will choose & configure a production incident reporting tool. Once robots
have been created & are being used in production, if anything goes wrong the
users need a way to report error quickly & get them fixed as soon as possible.
• They will select & execute a development methodology .(Agile/Waterfall) As
mentioned earlier RPA is best done using agile workflow, but it is certainly
possible to achieve it with waterfall aswell. It comes down to your enterprise
culture.
• They will publish scripting conventions & standards, & somebody on the team will
also likely be reviewing the scripts that comprise the robots.
• Some role within the RPA CoE will manage virtual workers, passwords, & so on.
And of course in doing so they will leverage IT provisioning & entitlement.
• CoE will build test, monitor & fix robots
• They will plan & manage upgrades & testing of RPA tools in the orchestration
server.
 • Finally they will establish, measure and report metrics, because at the end of the
day, the robots are being created to satisfy business needs. So it is important that
the robots emit telemetry that will help us generate daily, weekly, or monthly
metrics so the leadership team can see that the investments in robotics is paying
off.
CoE Composition:
• If its was 100% business people,
o you’d have lots of business process knowledge,
o but you might never maximize the RPA tool capabilities.
o You’ll probably have a lot of script or workflow duplication
o & automation would likely block manual work since it would be running
locally on the user’s computer
o There is potential for data loss because business folks don’t necessarily
understand how to use version control tools to manage scripting
• If CoE was 100% IT,
o You would wind up with effective tool & infrastructure skills,
o But there would likely be a lot of time spent communicating business
processes & requirements in order to build the robots
• Ideally the RPA CoE should be a hybrid model
o We should synergize IT resources with business resources.
o Each of these departments has complimentary skills for developing robots
o After robots have been created & they’re in production, if a production
issue occurs, sometimes it requires both business and IT to diagnose & fix
problems.
o Favorite matra for the RPA endeavor is that it’s business led & IT-enabled

RPA CoE roles

• Sponsor
• Champion
• Steering Committee
 • Change Manager
• IT lead
o Solution Architect
o Robot developer
o Robot Tester
o Infrastructure Engineer CoE Members
• Business lead
o Business Analyst/UAT
• Virtual workforce supervisor
• Robot service & Support

Depending upon the company’s culture, it might make perfect sense to also include a
change manager within the CoE, as discussed before change management initiatives
should happen early & often so that you can garner the most support for your RPA
efforts.
CoE Operations
Once CoE is formed, below is the roadmap for day-in & day-out operation.
Below is the high-level representation of the tasks discussed before.

• COE should choose processes wisely

• They will execute the demand management process
• Build, test, monitor & fix robots
o Of course, adhering to the chosen development methodology
o They will manage the virtual workers including passwords & entitlements
o They will definitely follow scripting conventions & standard that were
developed by the COE
• They will execute the production incident process whenever it is required
o There needs to be 24/7 monitoring
o & ideally you can incorporate some alerting technology like xMatters that
will first send emails to the on-call resources, and if emails are not
answered it will send text messages. & it text messages are not answered,
it will escalate to other people on the call chain until the problem is
resolved.
o You will be capturing both the operational & failure metrics.
 o Always prepare for DR (disaster recovery) events. If some reason the server
fails, you’ll have to have the appropriate measures in place to detect &
resurrect any servers that RPA endeavor relies on.
• Plan & excute any tool & infrastructure updates necessary. & this will have to
happen in the test environment, the stage environment, & the production
environment
• Measure & report metrics to the senior leadership team
• Maintain frequent consistent communication with the steering committee to
ensure the RPA effort is causing the desired business outcome.

Hub & Spoke

Creating an RPA CoE will definitely get you started down the right path in your RPA
endeavor. There’s a chance though, if you have a large company with many different
departments that have diverse automation needs, that your CoE simply won’t be able
scale to handle the needs. In that case, in the future you’ll want to consider a hub &
spoke CoE model where you have a main CoE that centralizes things that can be
centralized among the various business units or divisions & then your divisions or
departments like HR, Sales, accounting & marketing will each have their own embedded
CoE, which has more specific people, process & technology that’s most relevant to that
division or department

HR

Marketing Main Sales

Accounting
Summary:

• CoE is simply organizational construct to ensure RPA is successful & efficient

• It is a single point of accountability
• The ideal roles are:
o Change Manager
o IT lead
▪ Solution Architect
▪ Robot developer
▪ Robot tester
▪ Infrastructure engineer
o Business lead
▪ Business analyst/UAT
o Virtual workforce supervisor
o Robot service & support

Tips for RPA success

• When you begin your RPA program make sure you have clear strategy for the use
of RPA across the enterprise.
o Having broad view like this can help with things like tool selection & even
producing reusable workflows.
o Make sure you link the RPA effort to strategic senior leadership team
imperatives, such as increasing efficiency & increasing agility. Senior
leaders are all about maximizing the profit margin, & that comes back to
efficiency & agility.
• Be sure to involve IT from the start.
▪ While RPA use within & is managed by business, it’s still governed
by IT practices
▪ Software robots all require a PC, a user account(s), & access to one
or more applications, & IT is the team that most often provides that
kind of stuff.
▪ You’ll also eventually need development & stage environments &
most likely you’ll need test data in those development
environments as well, so you can develop & test your robot outside
of production.
 ▪ In addition to all that, real time IT support is definitely going to be
needed when your robots are being used in production, so involving
them early will allow them to get on board & build in the necessary
processes.
• Be sure to choose the most sensible process candidates
o We are talking about repetitive, high-volume, rule based process
o That are standardized & have structured Input & output
o Make sure to watch for opportunities to optimize the process along the
way
• Have a people plan early. This is one aspect of change management. People will
definitely start to get nervous when terms like software robots are thrown
around, so make sure you communicate early & often to assuage some of those
fears & help people realize that software robot is a good thing.
• Be prepared for hidden costs
o IT infrastructure such as databases, physical & virtual machines,
o & IT resources time to get RPA up & running are absolutely costs that you
need to factor in.
o If you decide to use a vendor then consultancy costs from partner
companies could come along
o Some new roles maybe created for RPA that add additional salary cost as
well.
• Think about scale early & often. Scaling means things like having good processes,
which includes
o Software version control, so robot developer can more easily merge their
changes & not worry about deleting others developer’s changes.
o Creating reusable workflows, so different developers & different robots
aren’t repeating the same work in different ways.
o Automated robot agent installations, when you are creating attended
robots, the robot agent needs to be installed on the humans agent’s
machines, & if you have 10, 20 or 100 or 1000 users you’re not going to
want to be doing those installations manually.
o Infrastructure updates & rollbacks. Periodically you’ll have to update your
orchestration server, your developer IDEs & even the robot agents. So
having a schedule for when that occurs & rollback plan if things go badly is
going to be important.
o Plan to supervise, manage, & maintain a virtual workforce. Its rare that 10
or 20 people on the same team would call in sick on the same day. But if
 10/20 of your software robots tips over, people are going to notice quickly,
& you’ll need to have a plan to get them back online.
• Centralize wherever it makes sense to do so.
o Instead of bunch of different business teams buying their own tools &
getting license individually, if one centralized group purchases the tools &
licenses on behalf of the whole organization you’ll certainly get better
prices on your licenses & be able to manage them more effectively.
o There will be less redundancy, as mentioned above if people communicate
about the kinds of robots they’re building & the dependent systems & you
are able to create one set of reusable workflows, that’s much better than
having different teams creating workflows for the same systems over &
over again
o You’ll get greater insights about processes & business activity if everybody
shares information.
• Set realistic expectations
o RPA is a tool, not the tool. It should be one part of a wider technology
strategy. A big part of that strategy is going to include AI in the near future.
o RPA can’t do everything yet. For instance, it challenging for software robots
to read handwritten language/documents. But there are vendors who are
working hard to solve this problem.
o Bear in mind there may still be a need for human intervention to manage
exceptions
▪ Business rule exceptions will occur on occasions
▪ Sometime the applications your robot depends on won’t respond as
expected. So whether there is a backup human to do this process or
a human to monitor & repair the robot, you’ll definitely still going to
need some humans around.
• RPA is more than just software robots
o You have got planning & governance
o Process improvements
o Sliding scale of human efforts v/s IT integration v/s RPA & that choice b/w
attended & unattended robots.
o We have got exception handling
o Change management the composition of the CoE & the variety of other
things.
NEXT STEP
For Business leader/Analyst
• Consider academy.uipath.com course
o RPA implementation
o RPA Analyst
These provide high level overview from UiPath perspective & gives some hands on tool
that would help with process selection & prioritization & different aspects of how to
execute an RPA project.