Published by:

A big part of business continuity is to provide a constant stream of potential projects. This task is called the “Discovery Process.” RPA is no different. In RPA’s context, you will need to discover those new processes. Once a Center of Excellence (CoE) is up and running, the most significant goal they will have is to propose and keep finding processes to automate. Read on to see how the discovery process works in RPA.

Most Companies Follow a Natural Process Discovery

When Companies start their RPA journey, they may already have an idea of the typical processes they want to automate. Those processes are usually the ones that cause some friction to its users, between departments, or are just visible low-hanging-fruit. When a Company reaches somewhere around their first 50 to 100 processes automated, they start to slow down the rate of automation.

By this time, it is critical to have a CoE and a proper discovery methodology in place. Having one will assure you that you can still discover more processes. The next set of processes will be critical to your CoE growth. These processes will be those that were not as obvious candidates for automation.

We know that about 80% of automatable processes are left undiscovered. It would be a huge missed opportunity to have only 20% of your possible automation in place. Think about how much ROI Dollars you would be leaving on the table!

Follow a Process for the Discovery Process

Understanding your business processes will take you about 65% of your automation efforts. The more you can simplify these steps, the more ROI you can bring back into your organization. It is critical to have a methodology, a process to follow, and the tool to do this.

Most Companies Hit a Discovery Process Slowdown

You have reached that plateau at around 50 to 100 processes automated. If you know what comes next, and how to tackle your challenge, you are all set. You need to understand the way to discover new processes so you can go past this plateau. Regardless of how you will find those processes, knowing how it all works will help you understand and take better advantage of those tools.

There are three tools for you to discover processes.

The Discovery Workshop

The first one is the guided methodology itself. This process involves a few prep-type meetings and culminates with a workshop meeting. The final output from these steps is a list that includes all the essential information you would want to plug into a decision matrix that the CoE can use as their backlog.

Process Mining and Process Mining Tools

The second one is the act of looking at logs to infer what processes are completing while the applications run. Your process mining tool will look into an application or system log. The mining tool will be able to infer a list of processes that were running on the device. Process mining tools are somewhat successful in achieving their goals, but they have a high dependency on the quality of the logs to analyze. If an application generates incomplete logs or logs process documentation sparingly, then the mining tool cannot infer as accurately as expected.

Discovery Tools

The third tool involves an application that “observes” what the manual process does and outputs the process document specs. In some cases, they output more complex content, such as the code for the automation bot itself. You could say that the Discovery Process tools are the automatic-equivalent to the Discovery Workshop process.

The discovery tool looks into the activities that the manual user performs and then outputs documentation that describes the process. Note that this is no different than the manual process of analyzing before automating such a process with your RPA platform. When employing such a tool, you are speeding the documentation that a Business Analyst would create. 65% of the time spent in understanding the processes mentioned above can be reduced considerably just by using this type of tool. How convenient would it be if there was a discovery tool that can output a spec document for the process you need to automate. Some of the discovery tools, like Discovery Bot from Automation Anywhere, can do that and even create the initial bot structure in the development environment as an output from the discovery tool!

We are automating the process of automation!

While discovery tools are great, let’s also be realistic; you are not avoiding the whole development process. The bot that comes out automatically of the discovery tool is just a starter-bot that has the basic structure you would get from a business user after having created a basic bot at best. It is not a complete bot that follows best practices and all, but a great result nonetheless. It is just another task in the list that gets completed faster, thanks to RPA.