MON

Monitoring & Drift

Operational controls for monitoring & drift, with maturity levels, evidence requirements, and implementation guidance.

Not sure where to start? Answer 3 questions and get a tailored compliance action plan.

What applies to me? →

Filter:Low effort Medium effort High effort

Agent-relevant

Clear

1 control matching filters

MON-006

Agenthigh

Behavioral Anomaly Detection for Agentic Systems

Implement monitoring that detects when AI agents deviate from their expected behavioral envelope, unusual action sequences, unexpected resource access, or goal-directed behavior inconsistent with assigned tasks.

Monitoring & Drift, tracked weekly

New monitoring & drift controls and the regulatory developments driving them, plus everything else changing in AI governance. Every Thursday.