A Digital Twin for Business Processes
A research-grade engine that takes a business process from historical event logs to a live, predictive twin, built in three validated maturity levels: digital model (discovery, conformance, object-centric process mining), digital shadow (streaming sync, drift detection, prediction), and digital twin (what-if simulation, optimization, and a closed feedback loop). Validated end-to-end on the BPI Challenge 2017 loan-application log: 31.5k cases, 1.2M events.
- Digital Model - process discovery, conformance & object-centric mining on 1.2M events
- Digital Shadow - streaming sync, concept-drift detection & ML prediction (7.13 d MAE)
- Digital Twin - calibrated what-if simulation, within 7% of real cycle time
- Closed loop: observe, forecast, detect divergence, recommend from live state


















