Tutorial 1: AI-Assisted Business Process Monitoring

Andreas Metzger

Business process monitoring involves tracking and analyzing operational business processes to gain insights into their performance, identify bottlenecks, and facilitate that they are running efficiently. This tutorial introduces the participants into how modern AI methods can be employed to realize predictive as well as prescriptive business processes monitoring. Where predictive monitoring helps to answer “what will happen and when?”, prescriptive monitoring allows answering “when to intervene and how?” Together, these monitoring approaches assist process managers and operators in deciding on when and how to intervene during an ongoing business process in order to prevent or mitigate the occurrence of an undesired process outcome. The tutorial introduces the participants to advanced deep learning methods for business process monitoring: deep supervised learning for predictive monitoring, and deep reinforcement learning for prescriptive monitoring. The tutorial positions these deep learning methods within the overall framework of business process monitoring systems and explains how deep learning helps to address key challenges. It presents empirical results on the effectiveness and cost savings of these deep learning methods, which are distilled into a set of recommendations for selecting appropriate deep learning methods in practice. Finally, the tutorial provides an outlook on future directions in AI-assisted business process monitoring, particularly elaborating the opportunities introduced by large language models (LLMs) and the need for explainable AI (XAI).

Tutorial 2: Constraint-based reasoning and analysis for BPM: CSP to the rescue

Alessandro Gianola, Andrey Rivkin and Mateusz Slazynski

Formal methods have always been an integral part of the BPM lifecycle. They are mathematically grounded techniques used to specify and analyze complex systems with a high degree of precision. In BPM, they help ensure that processes are correctly specified and capable of achieving the intended outcomes, by detecting design flaws and verifying compliance with defined goals. Among these techniques, the Constraint Satisfaction Problem (CSP) stands out as a powerful approach to specifying and solving problems with clearly defined constraints. CSP techniques are known for their high performance, yet a major barrier to broader use lies in the difficulty of defining suitable encodings, which require expert knowledge. Nevertheless, modern tools have made substantial progress in improving CSP accessibility for non-experts. In this tutorial, we define CSP, motivate its use in BPM, and showcase how two CSP instances, one over Boolean and the other over structured domains, can be used to solve BPM analysis problems via suitable encodings. Such encodings will be demonstrated during hands-on sessions with the help of state-of-the-art CSP tools. As a key learning outcome, participants will gain a deeper understanding of CSP-based formal methods and their integration into process analysis.

Tutorial 3: Business Process Optimization

Remco Dijkman and Arik Senderovich

Business Process Optimization (BPO) concerns making optimal operational decisions during process redesign or process execution—such as task assignment, resource allocation, resource staffing, and case advancement (e.g., admitting a patient). These decisions must balance objectives like cost, time, and satisfaction while respecting constraints like availability, deadlines, and budgets often under uncertaitny. Though related to Prescriptive Process Monitoring (PrPM), BPO addresses constrained combinatorial problems that often exceed real-time capabilities. This tutorial introduces key BPO problem types and solution approaches, illustrating specific problem-solution pairs that were already explored. While not exhaustive, the classification aims to provide a structured overview and spark future research in this emerging area.