Using Logs to Reduce the Impact of Process Variability and Dependence on Practitioners in Requirements Engineering for Traditional Business Process Automation Software

Context: Business Process Automation (BPA) is adopted by organizations to improve efficiency, reduce costs, and increase overall business performance. Traditional Business Process Automation (TBPA) is one of the main approaches employed to develop a BPA. TBPA entails the development of BPA in a programming language to integrate relevant applications in the digital ecosystem to execute a given process. Process variability and practitioner unavailability are issues that encumber the requirements engineering for TBPA software. Objective: This work proposes Requirements with Logs (RWL), a log-based approach for TBPA software to reduce the impact of these issues, by providing a higher alignment among business process requirements and software architecture, and employing process mining to semi-automatically discover the business process during requirements elicitation. Method: The research conducted a case study in a technology institute to assess RWL and report its results in practice. Results: The results revealed significant improvements in adaptability to business process changes, in time spent with practitioners, and in software development. RWL also presented limitations, including human intervention to accurately obtain the business process, complexity to trace the process into the architecture, data privacy concerns, and risk of network traffic overload. Conclusion: This research demonstrated the effectiveness of RWL to minimize the impact of process variability and the dependence on practitioners.