AI-Augmented Business Process 

Management Systems

In this talk, I will present a vision of AI-Augmented Business Process Management Systems (ABPMS): process-aware information systems that sense, reason, and act within a human-set frame to continuously adapt and improve processes, in an explainable and conversational manner. In such systems, execution flows are not entirely pre-determined by models or rules, adaptations may occur without explicit changes to supporting software systems, and improvement opportunities may be autonomously discovered, validated, and enacted by AI agents. I will position an ABPMS as a pyramid of capabilities, with descriptive, predictive, and prescriptive analytics forming the foundation layers that support AI-augmented process execution. I will sketch an ABPMS architecture that brings together various technological building blocks (process mining, simulation, causal policy learning, LLM-based agents) to enable autonomous process execution and improvement within a frame of constraints and goals. I will also map the scope of processes supported by an ABPMS within a triangle of hybrid processes, spanning from mostly manual processes at one base vertex, through automated (rule-based) workflows at the other base vertex, up to fully autonomous (agentic) processes at the apex.