5. A Comprehensive Definition to Robotics Process Automation
A software technology that enable us to create rules and logics to effectively mimic a user's action on a computer thereby
interacting with different software systems at the GUI (graphical user interface) or presentation layer, the same level as a
“human” user of the system.
Emulates human execution of repetitive process with existing business
application.
Virtual workforce controlled and monitored by business operation
team.
It can be trained business users via configuration.
Operate 24x7
Sit alongside existing infrastructure governed, controlled by IT.
Generate audit logs.
Easy to deploy.
5
7. Dual data entry scenarios:
•Data manually entered in
one system need not be
reentered manually in any
other system. RA replaces
such dual human effort
since invoices are indexed
in the workflow and then
manually reentered in ERP.
Straight-through processing:
•Inputs arriving from various
systems such as web pages
for customer orders,
workflow for invoices,
emails, or Excel files must
be entered into ERP.
However, if the input is
clean and the rules are well
laid out, that data entry can
be done through RA.
Virtual “integration”
between different systems:
•Standalone, legacy, ERP, or
workflow systems often
don’t “talk” to each other,
and integrating them would
cost millions of dollars and
precious IT time. RA can
provide lightweight
integration connecting
disparate systems at the
user interface level.
Responding to data
extraction and reporting
requests:
•When data and report
requests come from
multiple process owners,
vendors, and even end
customers, employees log
into a system to extract the
data, format it, and send an
email to the requestor. RA
lends itself very well to
such rule based tasks.
Rule-based decision-making:
•RA can execute decision-
based tasks provided the
rules driving those
decisions are well laid out.
For instance, on an invoice
coming from a utility
vendor, RA can change
payment terms to
“immediate” from
whatever is on the invoice.
If an order is above
Best use
7
8. Key things to identify scope
8
•Total FTE count in the process
•Monthly volume of work being
done
•Lead time involved
•Takt time
Human Factor
•Development time is
proportional to complexity
involved in the process.
•The more complex is the
process more time it will take
to develop.
•Development cost will be
involved
•More technical issue will pop
up.
Complexity
•Process stability determines
longevity for Automation
•Win-Win case if process change
is minor and requires fewer
code tweaks.
•It will add up to re-work cost if
process change is major and
requires reengineering.
•Don't take up process for
automation that calls for
frequent changes.
Process Stability
9. Key things to identify scope (Contd.)
9
•Inputs could be structured or
unstructured
•For unstructured input, there is
a need to standardize the input
format.
•Robot feeds well on structured
input rather than unstructured.
•It depends on operation if they
welcome this change for the
robots.
Input Type
•Automation works best on Rule
based process as compared to
those with high number of
decision making.
•Cognitive capabilities are still a
challenge for Automation tools
in the market.
•Go with a process that is
majorly rule based and involves
lesser decision making and
judgments.
Rule based vs. Decision based
•A good number of human errors
in a process can be useful from
Automation standpoint.
•Robots are efficient in cut-copy-
paste, data transfer, navigation
etc.
•Removing these overheads from
a process would account for
greater productivity, quick
changeover, reduction in cycle
time and speedy operation.
Error Rate
10. Discovery phase
•2-3 hours of meeting with
key business
stakeholders,process
owners and associates to
get high level overview to
identify suitable
candidates for RPA
Process Assessment
•1-2 Days workshop with
process owners and SME
to understand the As-is
process for automation
•Process maps
Business case
proposal
•Evaluate top RPA vendors
•Calculate RPA benefits
such as Cost , Time , FTE
•Defining project cost
•Approvals
10
Robotics Process Automation Planning
Planning phase takes around ~3 or 4 Weeks
11. Plan/Analyze
Identify
ImplementTest
Go-Live
Lifecycle
Identify tasks that can be automated
easily, start with smaller tasks, repetitive,
a pain point for
existing manpower, not too complicated,
task that have buy-in from the team for
automation, short
implementation timeline, quick result.
Start by targeting processes that will benefit the organization and
stakeholders the most.
Look for highest ROI and preferably start small, short implementation
timeframe, quick results.
Create work allocation tracker for the development
team. Assign process discovered during the early phases
to the individual RPA developer. Create uniform
architecture for the developers to follow, ensure best
practice while developing the solution.
Strict adherence to development timelines,
performance and quality. Perform a thorough
functional and unit testing.
Post implementation of RPA, thorough testing is required, to
check for production quality, simulate exception handling,
compare existing turnaround time and quality, obtain UAT
signoff, run RPA in parallel with existing process for a couple
of months,
check overall performance before go-live.
Post successful parallel run, check overall results, ensure
existing manpower is available during the post go-live
stabilization phase, obtain required signoff
, setup a COE, training plan for operation team for future
troubleshooting and bug fixing, realign FTE to more
challenging task, enhance there skills set, prepare a case
study
1.
2.
3.
4.
5.
11
12. 12
0
25
50
75
100
Low Medium High
Timescale(Days)
Process Complexity
RPA Timeframe
1. Process Analysis and Solution design.
2. Development and testing
3. Move to production
Process finished in 3 weeks against
assumption of ~1 week
Process finished in 6 weeks against
assumption of ~2 week
Process finished in 8 weeks against
assumption of ~3 week
RPA Timeframe
1. Process Analysis and Solution design.
2. Development and testing
3. Move to production
1. Process Analysis and Solution design.
2. Development and testing
3. Move to production
13. SWOT
Strengths
• Fully/Partially automated process
• On time Delivery/Customer satisfaction
• Time Saving
• Cost Reduction
• Increased ROI
• Audit Logs to support compliance
• Secured Access
• Rapid Development
• No change to business application
• Continuous improvement
• Easy to deploy and re-configure
• Increase productivity by eliminating wastes.
Opportunities
• RPA learning for Ops
• Process improvement
• Strong case to attract other business
• Encourages process standardisation across
different business areas for the same tasks
• Leverage FTEs to other high skilled jobs
• Gain share
• Roadmap for future projects
• Building COE
• Strong BCP
• Incorporating NLP and Data quality
• Change Management
Threats
• Cognitive process
• Application downtime
• System/Hard disk crash
• Major change request
• Business process pulls out
• Unstable process
• Decreased timelines for development team
• Upgrades to underlining application poses a
challenge.
• Consideration on existing systems capacity and
SLA provisions,
• Scalability of RPA Centre of Excellence across a
breadth of business functions
Weaknesses
• 5% -10% error rate
• Quality check
• Coding standard
• Performance variance across system
• Exceptions
• Unstructured input for processing
• Limited support for browsers such as chrome,
Firefox etc.
• Dependency on NLP tools reading and
understanding free flow texts.
S W
TO
14. Building a unified view of
the customer
Improvement in customer
satisfaction index
Up to 80% reduction in
AHT
Up to 90% reduction in
ART
No impact on the security
as RPA only does what
agent does.
Increased ROI within 3
months (Avg)
Same headcount for triple
the activity.
Benefits
14