Survey of Explainability within Process Mining: A case study of BPI challenge 2020

The need for explainability in Business process management is tremendously increasing, especially in the age of generative AI. The number of published articles on explainable AI (XAI) has skyrocketed for five years. AI impacts the decision-making process in business analytics. Process mining as a sub-discipline of data science can play a role in explainable business decision-making. Process mining exhibits its intention in process discovery, performance measures of processes, and process improvements based on the event logs. Although the accuracy of the outcome of process mining models has been investigated at a certain level, the explainability of those is possible through the discretization of the analytic steps. As an initial step in exploring the explainability of process mining, this research conducts a technical analysis of 37 research papers submitted to the Business Process Intelligence (BPI) Challenge 2020. The main focus of this analysis aims to answer the question, "How and why a process model is produced? " To make a foundation for the research question, the notion of explainability is explored based on an Explainable AI ontology. Due to the small sample size, the study cannot identify clear trends of explainability in process mining. However, the results conclude that explainability depends on the process model’s transparency and reproducibility. Moreover, further research with a large sample size is required to understand the discrete factors impacting decision-making in business process management.