What is process mining?

Definition

Process mining is an analytical discipline that uses event log data from IT systems to discover, monitor, and optimize business processes. The field was pioneered by Wil van der Aalst at Eindhoven University of Technology in the late 1990s. By analyzing timestamps and activities recorded in ERP, CRM, and workflow systems, process mining software reconstructs actual process flows as they occur in practice — not as they are documented on paper. This reveals bottlenecks, deviations, compliance violations, and automation opportunities that are invisible in traditional process analysis.

Three Types of Process Mining

Data Requirements

Process mining requires event logs with three minimum fields:

• Case ID: A unique identifier linking all events belonging to one process instance (e.g., order number, ticket ID)
• Activity name: The specific step or action performed (e.g., "Invoice Received," "Payment Approved")
• Timestamp: When the activity occurred

Additional fields that enrich analysis include: resource (who performed the activity), cost, department, and custom attributes. The quality and completeness of event logs directly determines the quality of process mining output.

Connection to Automation

Process mining connects to automation in three ways:

1. Discovery: Identifies which processes are the best candidates for automation based on volume, repetitiveness, and standardization
2. ROI measurement: Quantifies the before-and-after impact of automation by comparing process metrics (cycle time, error rate, cost) pre- and post-automation
3. Monitoring: Provides ongoing surveillance of automated processes to detect exceptions, performance degradation, and drift from expected behavior

Key Vendors (as of 2024)

The process mining market reached approximately $1.4 billion in 2024, growing at an estimated 40% compound annual growth rate.

Task Mining

Task mining is a related discipline that records desktop-level user activities (mouse clicks, keystrokes, application switches) to discover user-level process steps that are invisible in system event logs. Task mining complements process mining by filling the gap between system-recorded events: it reveals what humans do between the timestamps that appear in ERP and CRM logs. UiPath, Celonis, and Automation Anywhere all offer task mining capabilities.

Limitations

• Data quality dependency: Process mining output is only as reliable as the event logs it ingests. Incomplete logging, inconsistent activity naming, and missing timestamps degrade results.
• Privacy concerns: Event logs may contain personally identifiable information (employee names, customer data) that requires anonymization before analysis.
• Multi-system complexity: Processes that span multiple IT systems require event log correlation across different data formats, timestamps, and case identifiers.