When an organization is considering the implementation of a new technology, they want to identify the benefits before making a large investment. Therefore, running a pilot program is extremely important as it allows an organization to test the solution on a small scale and decide if the initiative suits their needs. Keep in mind, a pilot is different than a proof of concept (POC) but more on that another day.
Anyway, choosing the best pilot process is the most important part of a successful robotics operation. Many companies want to begin with automating their pain points, however, in many cases this is not the best place to start for a pilot process. When doing a pilot, you want to ensure the process you choose will be able to demonstrate the value of RPA and not select a project that will bury you in development or process re-engineering. In this article we will outline seven key features of a good candidate for a pilot.
Start small
A good way to start an RPA project is to choose a smaller process to build up knowledge and required capabilities before scaling up. However, the pilot process cannot be too simple, remember that it should be relevant and bring benefits to your organization.
Keep it simple
Avoid selecting processes that are too complex, or ones with multiple paths, look for something simple, where the happy path covers at least 80%. Complicated processes increase the risk of delays due to the different types of exceptions. Such processes usually have several people involved, which may make it too difficult to map each step properly. A better choice would be to select a process that requires only one subject matter expert to described in detail.
Minimize the level of effort
Another important factor is ease of automation. Although this may be difficult to determine, especially at the beginning of an RPA journey. An experienced RPA lead developer can perform an application assessment to estimate the level of effort required. They are familiar with the capability of the tool and can identify potential issues and present possible solutions. Typically processes where we need to interact with virtual environments, such as Citrix, are not suitable for pilot projects as they need to utilize surface automation techniques. Instead choose easier to automate applications such as browser or windows-based applications.
Use digital and structured data
Processes chosen for the automation project should have input data supplied in digital and structured form. If the data does not come directly from the target application or system, consider creating an example excel template to simplify the ingestion of the data for the digital worker. Look for processes with structured data inputs, for example, XLS, XML, CSV, JSON. Also, avoid processes that rely on data from scanned documents!
Ensure processes follows strict rules
Avoid picking processes with poorly defined/documented rules and procedures or ones that require intuition. The RPA developer requires precisely described procedures, down to the click and type level, to train the digital worker. Finally remember to document accurately each step of the process, provide input data format and list all types of business exceptions.
Maximize the number of available test cases
To maximize the chances for a successful RPA solution, test the process with enough test cases to cover as many production scenarios as possible. In the User Acceptance Testing (UAT) phase ensure that test plans accounts for at least one full day’s worth of data. Running these tests properly will drastically limit any surprises in production
Identify the business value
As mentioned before, a pilot process should be rather small and simple, however keep in mind it’s relevance. Look for direct benefits for the organization such as employee optimization, decrease of throughput time, or reduction of errors. Avoid selecting any critical process, as risk of any mistakes could be unacceptable. Target no-risk or low-risk processes, keep the human in the loop having them make final decisions, and ensure the benefits are quantifiable with minimal effort.
Summary
There are no hard rules on how to choose a pilot project, however, following these tips should make life a bit easier and drastically increase the chances of success for an RPA initiative. A digital worker is similar to a new employee. Just as one would train a new employee to perform a task, a digital worker also requires a similar approach to such training. Not to oversimplify things, but when looking at potential pilot projects ask, “Is this task something we would allow a new employee to do in their first couple weeks?” If not, chances are there is a reason why that process might not be a best choice to kick off an RPA initiative! Additionally, consider working with a partner like us! Our experience with looking at potential processes, identifying their automation complexity, and estimating the level of effort required to implement. Plus, after the assessment, we have resources available to assist you further, or even complete the project from start to finish 😊.