Term

Process Mining

Process Mining is the analytical technique of reconstructing the actual flow of a business process from event-log data and comparing it to the documented one. An event log is three columns: case ID, activity name, timestamp. Feed enough in and the technique discovers the process as it really runs. Distinct from Business Process Management (BPM) software, which describes the intended process, and from Workflow Automation, which executes one. Process Mining looks at what happened.
Reviewed by 7wData

Why it matters

Process Mining is the answer to “we documented the process but nobody knows what really happens.” It exposes deviations, bottlenecks, rework loops, and shadow paths invisible to the process owner, who is looking at the diagram, not the log. For AI work the relevance is direct: agentic systems generate exactly the event-log shape Process Mining needs (case ID is the agent run, activities are tool calls, timestamps are free), so agent behaviour can be audited the same way human-driven processes are.

Where you’ll encounter it

Three contexts. A finance team uses it to find compliance deviations in Purchase-to-Pay or Order-to-Cash flows, the classic entry use case. A vendor (Celonis, UiPath, IBM, ABBYY, Apromore) sells it as a Continuous Process Improvement platform with a connector library on top. An audit board asks for evidence of actual-versus-designed conformance, and the answer is a conformance diagram, not a slide. One warning: garbage event logs produce garbage discoveries. The event-extraction step (what counts as a case, what as an activity, how to handle missing timestamps) is the make-or-break, and most failed deployments fail there.

Part of the 7wData AI Glossary. Tracking how concepts like this move in the expert conversation: daily signals at ins7ghts.com.