Model Risk & Governance

Data on This Page

Five production models owned by other agents (Structuring Detection v4.2, Entity Match Confidence Model, Fraud Risk Scoring v7.1, Sanctions Fuzzy-Match Engine, Network Anomaly Scoring), each with its owning agent, days since last validation, a drift score, and a status of healthy, drift detected or revalidation due.

What This Agent Does

Provides the model risk governance layer across the whole platform, tracking every production detection model regardless of which agent owns it, measuring statistical drift against its validation baseline, and flagging models overdue for revalidation so the platform stays audit-ready at all times.

Worked Examples

• Network Anomaly Scoring hasn't been validated in 61 days and has drifted to a score of 21 — flagged revalidation due, the most overdue model.
• Fraud Risk Scoring v7.1 shows a drift score of 14 after 38 days without validation — flagged drift detected.
• The Sanctions Fuzzy-Match Engine, validated just 9 days ago with a drift score of 1, remains healthy.