Mule graph detection
Network-graph analysis surfaces mule-account rings — fan-in/fan-out money-movement patterns that look unremarkable transaction-by-transaction but are obvious once graphed.
Self-learning feedback loop
Every analyst decision (confirmed mule, false positive) feeds back into the model via continuous backtesting — detection improves from real outcomes, not just periodic manual retraining.
How it fits with Device Intelligence
The same device and user graph that powers Device Intelligence’s device-to-user mapping is available to AML investigations — an account flagged for mule activity that also shares a device with several other flagged accounts is a materially stronger case than either signal alone.Who uses this
AML/CFT analysts
Work cases in the investigation console — graph view, transaction timeline, and evidence collection in one place.
Compliance officers
Review case dispositions, export evidence for regulatory filings (STR/SAR-equivalent), and track model performance over time.