Democratizing Robotic Process Mining: A Conceptual Framework for User Actions, Tasks, and RPA Bots

While Robotic Process Automation (RPA) enables companies to automate tasks in a comparable lightweight manner, the creation of such automations still requires a considerable manual effort. Robotic Process Mining (RPM) tries to mitigate this effort by analyzing recorded user interactions with a computer system to identify routines automatable by RPA and to derive corresponding RPA scripts. The rapid evolution of both RPA and RPM has led to ambiguities in the terminology used across research, prototypes, and industry. In this paper, we survey and structure the possible types of user interactions that can be recorded by analyzing the literature on RPM and the related field of user behavior mining. Additionally, we aim to foster a unified understanding of the key terms and concepts in RPM by proposing a conceptualization that draws upon ontologies of software, user interfaces, and RPA to embed this comparably new research area. To explore the practical applicability of the conceptualization, we conclude with a number of concrete usage scenarios that strengthen the design and development of RPA bots.