Close

Choosing the Right
RPA Pilot Process

August 2020 | Article

By Bhavik Patel

By Bhavik Patel

8 Tips for Choosing the Right RPA Pilot Process

As you’ve probably noticed, there has been a massive surge in robotic process automation (RPA) over the last few years. According to Statista, 24% of large organizations have already adopted it.


Are you thinking about moving towards automation? By far the best way to discover whether RPA is suitable for your organization is to run a pilot program. After all, it’s a large investment so you don’t want to jump in with both feet.


In this article, we outline 8 tips for a successful RPA pilot.


But first… Here’s how NOT to do it

How NOT to run an RPA pilot

When running a pilot program, organizations often make 3 mistakes. They begin by automating:

• Too many processes at once
• Their most difficult, complex processes
• Their pain points

As a result, the project gets buried in development and process re-engineering. The pilot becomes complicated, it slows everything down and the true value of RPA is lost.


In order to demonstrate the benefits of RPA and make the right investment decision, there are better ways to go about it. Here are our 8 tips:

1. Start small

Starting with a relatively small process allows you to build knowledge and assess the required capabilities before scaling up.


Although small, these processes should nonetheless ease the journey towards your organization’s objectives. And, if the pilot brings tangible benefits to employees’ working life, all the better. It will encourage buy-in further down the line.

2. Low complexity, little effort

Your pilot should require little effort, so keep it simple. Automating
complex processes involves programming for a variety of exceptions,
and this causes delays.


The process `happy path’ – the default scenario with no exceptional
or error conditions – should cover at least 75% of the process.


When you’re thinking about processes for the pilot, we recommend
considering:


• The number of steps involved in the process
• How many applications it uses
• The number of business exceptions
• The structure and quality of data
• Volume of transactions

2. Low complexity, little effort

Your pilot should require little effort, so keep it simple. Automating complex processes involves programming for a variety of exceptions, and this causes delays.


The process `happy path’ – the default scenario with no exceptional or error conditions – should cover at least 75% of the process.


When you’re thinking about processes for the pilot, we recommend considering:

• The number of steps involved in the process
• How many applications it uses
• The number of business exceptions
• The structure and quality of data
• Volume of transactions

3. Standardized process

Whichever process you select, it should have well-defined and documented rules and procedures.


Always remember that the robot is only as good as the script you tell it to execute. In order to train the robot to follow the process as a human would, your RPA developer will need the process broken down at the keystroke level.


We recommend that at least one lead SME dictates the design for the robot. They perform the process manually every day so unlike the process owner, they’ll have in-depth knowledge of its challenges.

4. Simple integration

Bear in mind that RPA may not be the best option for some processes. For example, if it simply involves a defined transaction between two applications, API may achieve better results.


Most of the time, things aren’t that simple – but when you’re weighing up which processes to choose for the RPA pilot, it’s worth thinking about before making the final call.

5. Connectivity and stability

When you’re selecting a pilot process, choose a process that is working

with applications that are either windows-based or browser applications.

 

Avoid processes which involves interaction with a virtual environment.
It will overcomplicate the pilot because your developer will need to
use surface automation techniques.

 

Before making the decision, RPA developers should always carry out
an assessment of all the applications related to the process. Amongst
other things, they should look out for any stability issues, and make
sure the application is not due for an upgrade.

5. Connectivity and stability

When you’re selecting a pilot process, choose a process that is working with applications that are either windows-based or browser applications.

 

Avoid processes which involves interaction with a virtual environment. It will overcomplicate the pilot because your developer will need to use surface automation techniques.

 

Before making the decision, RPA developers should always carry out an assessment of all the applications related to the process. Amongst other things, they should look out for any stability issues, and make sure the application is not due for an upgrade.

6. Structured and digitized data

To keep things simple, stick to structured and digitized data. Steer clear of unstructured data, scanning documents or any information that is not digitized. You want to avoid having to integrate other AI technologies.


It’s fine for the input data to come directly from an application, or even an Excel file, but the data must be structured. That way, the robot knows exactly where to pick up information.


There are various methods for pulling information from PDFs, scanned documents and other files – but this will come later, not during the pilot.

7. Testing

Testing throughout the project is crucial. Although this can seem like a daunting prospect, thorough testing will always pay off.

 

Your developer should test each component of the RPA solution during configuration and then run end-to-end testing before handing it over to the business for UAT (user acceptance testing).
 
UAT is important because it’s an opportunity to break the solution. It reduces the chances of RPA breaking when it’s in production.
 

There should be enough test cases to cover the `happy path’ – plus tests to cover each exception to the process to make sure the robot can handle every scenario appropriately.

 

During deployment, we recommend starting small. Work on just a few transactions, validate the results and then ramp up incrementally to full deployment once you’re happy with each result.

8. Business value and setting clear goals

In many cases, organizations begin their automation journey because they’ve heard fantastic claims about ROI.

You will indeed achieve cost savings, but while it’s certainly worth considering for the pilot, this should not be the overriding factor at this stage. ROI comes into its own later, when you’re prioritizing automation projects.

The key consideration is what your organization hopes to achieve from automation. There is no right or wrong answer to this – but in our experience, apart from ROI and cost savings, some of the most common objectives are these:


• Better employee experience and improved employee satisfaction

• Higher quality output
• Improved risk compliance
• Improved revenue
• More innovation

Your RPA pilot will set the stage for an enterprise-grade program, so it’s important to prove success early. Selecting the right process will deliver a strong start to your automation journey.

For more information on how to begin the journey towards automation.

Share this post

Share on linkedin
Share on facebook
Share on email
Share on print