Speeding Up Robotic Process Automation through Behavioral Observation

#RPA, #Architecture, #Automation

In the last decade, a large number of Robotic Process Automation (RPA) projects have been conducted. In essence, these involve the use of software to mimic the behavior of human employees. Typically, most commercial RPA platforms (e.g., BluePrism or UiPatch) rely on a manual analysis of the project context to identify the candidate processes that are suitable for automation. This is a time-consuming task. It does not only involve the analysis of much documentation, it is also necessary to determine the expected frequency of cases, the exceptions that can take place, and various other attributes that are complex to estimate.

In this talk, we will explore the benefits of an alternative approach, which is completely agnostic with respect to the specific technology that is being used. Our starting point is to use log information, which captures various attributes of how humans behave. Specifically, we start out by monitoring the interactions between the employees and the information systems within the project context. We will first address how to transform that observed behavior (i.e., in the form of screen captures, clicks and keystrokes) into a meaningful user interface event log (UI log) by using image-similarity techniques. We will proceed to show how this UI log can be leveraged to automate the process analysis and design phases in an RPA project. Beyond this explanation, we will elaborate on the open challenges as well as the opportunities that the UI log provides for further stages of an RPA project, i.e., development, testing, and maintenance.