AI is revolutionizing industrial operations, but its effectiveness depends on understanding business processes, where process mining plays a crucial role in making AI-driven solutions practical and compliant. This tutorial at ECAI 2025 will equip AI researchers, data scientists, and industry practitioners with process mining techniques to integrate AI into real-world enterprise settings, enhancing efficiency and decision-making.

Summary

AI is transforming industrial operations, but its success hinges on understanding and optimizing business processes. Despite recent breakthroughs in generative AI, it remains difficult to apply AI and ML when facing data in enterprise information systems. In these settings, one needs to deal with dynamic end-to-end processes. Process mining serves as the lens through which AI can be effectively applied in enterprise settings, ensuring that AI-driven solutions address real-world performance and compliance challenges. Without this process-aware perspective, it is impossible to apply AI and ML.

This tutorial, part of the the 28th European Conference on Artificial Intelligence (ECAI 2025), shows how process mining unveils the hidden dynamics of operational processes, enabling mainstream AI and ML techniques. Data-driven approaches such as automated process discovery, conformance checking, and predictive analytics are explained using a range of examples. By detecting process-related problems such as a bottleneck or frequent deviation, process mining serves as the lens to address performance and compliance problems. These techniques are widely applied in manufacturing, finance, logistics, and healthcare. Through real-world examples, participants will gain a practical roadmap for integrating process mining into AI initiatives.

Designed for AI researchers, data scientists, and industry practitioners, this tutorial provides the essential tools to bridge the gap between AI and enterprise reality-transforming data into actionable intelligence for robust, responsible, and efficient AI adoption.

Context

Process mining has emerged as a pivotal discipline that bridges the gap between process science and data science, evolving significantly since its inception in the late 1990s. The discipline of process mining has been instrumental in addressing fundamental questions about actual vs. assumed processes, identifying bottlenecks and deviations, and predicting performance and conformance problems. Concurrently, amazing breakthroughs have been achieved in Artificial Intelligence (AI) and Machine Learning (ML). However, it is not so easy to apply AI and ML in an enterprise setting. Industrial applications tend to focus on a single task and not end-to-end processes. This tutorial delves into the symbiotic relationship between these domains, emphasizing how process mining serves as a conduit between data-centric analysis and process-centric understanding. The ultimate goal of the tutorial is to show that process mining provides the lens to apply AI and ML and improve processes in manufacturing, finance, logistics, and healthcare. Participants will get a good understanding of state-of-the-art process mining techniques and will be able to apply this knowledge with limited effort.

Participants will explore foundational concepts of process mining, including process discovery, conformance checking, and enhancement. Building upon these foundations, the tutorial will illustrate how AI and ML techniques can be tailored to enrich process mining endeavors. For instance, decision mining leverages ML to uncover decision rules within processes, while predictive analytics forecasts future process behaviors.

The session will also address challenges inherent in aligning AI models with process data, such as handling non-parametric distributions and ensuring model interpretability. By examining recent advancements, including the application of generative AI in process modeling, attendees will gain insights into the evolving landscape of process mining research.

Designed for AI researchers, data scientists, and industry practitioners, this tutorial offers a comprehensive roadmap to effectively integrate AI and ML techniques within process mining frameworks. Through real-world examples and interactive discussions, participants will be equipped to transform raw data into actionable intelligence, driving robust and efficient process improvements.

Detailed Outline

Despite advances in generative AI, applying AI and ML in enterprise systems remains challenging due to complex, dynamic process and data scattered over multiple systems and tables. Process mining bridges this gap, ensuring AI solutions tackle real-world performance and compliance issues. This tutorial at ECAI 2025 explores how process mining uncovers hidden operational dynamics, enabling AI-driven insights. Key techniques like automated process discovery, conformance checking, and predictive analytics will be demonstrated through real-world examples across manufacturing, finance, logistics, and healthcare.

These are the topics to be addressed in the half-day tutorial:

1. Introduction to Process Intelligence and Process Mining

   • Definition and significance
   • Historical context and evolution
   • Core components: event logs, cases, activities
   • Software tools
   • Industry adoption

2. Foundational Techniques in Process Mining

   • Process Discovery

     • Alpha algorithm & Heuristic miner
     • Inductive miner

   • Conformance Checking

     • Aligning event logs with process models
     • Detecting deviations and compliance issues

   • Process Enhancement

     • Performance analysis
     • Model augmentation with additional data

3. Interplay Between Process Mining and Machine Learning

   • Overview of AI and ML in the context of process analysis
   • Challenges in integrating ML with process data
   • Tailoring ML models for process mining applications

4. Advanced Topics

   • Decision Mining

     • Creating situation tables
     • Extracting decision rules from processes
     • Utilizing classifiers

   • Predictive Process Monitoring

     • Creating situation tables for performance questions
     • Forecasting future process states
     • Implementing predictive models for proactive interventions

   • Object-Centric Process Mining (OCPM)

     • Moving from case-centric to object-centric
     • Advantages of OCPM

   • Generative AI in Process Modeling

     • Leveraging large language models for process model generation
     • Case studies

5. Challenges and Considerations

   • Data quality and preprocessing
   • Ensuring model interpretability and trustworthiness
   • Addressing ethical and compliance concerns

6. Practical Applications and Case Studies

   • Real-world implementations in various industries
   • Lessons learned and best practices

7. Future Directions in Process Mining and AI Integration

   • Emerging research areas
   • Potential technological advancements
   • Open challenges and opportunities for innovation

The tutorial will provide ready-to-use data sets and open-source software. Participants are encouraged to apply the concepts introduced in this course. Next to providing Python-based software, the Celonis platform is used to illustrate industrial-scale applications.

Audience and Background

No prior knowledge of process mining is needed. The assumption is that participants have basic computer science skills and are familiar with common Artificial Intelligence (AI) and Machine Learning (ML) concepts. The topic aims to address a mixed audience. This includes:

• Introduce an established subdiscipline of AI/ML that participants may not be aware of.
• Provide new challenges for AI/ML experts, broadening the scope of applications.
• Trigger participants to apply and develop AI/ML techniques in industrial applications.
• Provide a compact overview of approaches connecting process science and data science.

About the Presenter

Wil van der Aalst is a full professor at RWTH Aachen University, leading the Process and Data Science (PADS) group. He is also the Chief Scientist at Celonis and part-time affiliated with the Fraunhofer-Institut für Angewandte Informationstechnik (FIT). Currently, he is also deputy CEO of the Internet of Production (IoP) Cluster of Excellence, co-director of the RWTH Center for Artificial Intelligence, and speaker of the RWTH ICT Profile Area. Until December 2017, he was the scientific director of the Data Science Center Eindhoven (DSC/e) and led the Architecture of Information Systems group at the Technische Universiteit Eindhoven (TU/e). He was also on the Board of Governors of Tilburg University from 2015 until 2023. Since 2003, he has been an adjunct professor at Queensland University of Technology.

He is commonly known as the “Godfather of Process Mining” and one of the leading experts in the field of data science, making him the ideal candidate for such a tutorial.

Wil van der Aalst has published more than 300 journal papers, 35 books (as author or editor), 725 refereed conference/workshop publications, and 90 book chapters. Many of his papers are highly cited (he is one of the most-cited computer scientists in the world and has an H-index of 184 according to Google Scholar with over 158,000 citations), and his ideas have influenced researchers, software developers, and standardization committees working on process support. According to Research.com, he is the second highest-ranked computer scientist in Germany and ranked 9th worldwide.

He is also an IFIP Fellow, IEEE Fellow, ACM Fellow, and elected member of the Royal Netherlands Academy of Arts and Sciences (Koninklijke Nederlandse Akademie van Wetenschappen), Royal Holland Society of Sciences and Humanities (Koninklijke Hollandsche Maatschappij der Wetenschappen), the Academy of Europe (Academia Europaea), the North Rhine-Westphalian Academy of Sciences, Humanities and the Arts (Nordrhein-Westfälische Akademie der Wissenschaften und der Künste), and the German Academy of Science and Engineering (Deutsche Akademie der Technikwissenschaften). In 2012, he received the degree of doctor honoris causa from Hasselt University in Belgium. In 2015, he was appointed as an honorary professor at the National Research University, Higher School of Economics in Moscow. In 2018, he was awarded an Alexander-von-Humboldt Professorship, Germany’s most valuable research award (five million euros).

For more information about his work visit: www.vdaalst.com, www.processmining.org, www.pads.rwth-aachen.de, and www.workflowpatterns.com.

Related material

• W.M.P. van der Aalst. Object-Centric Process Mining: Unraveling the Fabric of Real Processes. Mathematics, 11(12):2691, 2023.
• W.M.P. van der Aalst. Process Mining: Data Science in Action. Springer-Verlag, Berlin, 2016.
• W.M.P. van der Aalst and J. Carmona, editors. Process Mining Handbook, volume 448 of Lecture Notes in Business Information Processing. Springer-Verlag, Berlin, 2022.

Additional Resources