Monitor risky agents and keep agent governance current

Short summary

This article covers the continuous monitoring phase of AI agent governance using Microsoft Entra Agent ID, emphasizing that governance is an ongoing operational programme, not a one-time project. It outlines key monitoring areas including risky agents, sign-in logs, audit logs, owner/sponsor gaps, access package expiry, and Conditional Access impact. Recommended actions for high-risk agents range from review to disabling or retiring the agent identity.

• •Agent governance requires continuous monitoring — not a one-time setup project
• •Key monitoring areas: risky agents, sign-in/audit logs, owner gaps, access expiry, attribute drift
• •High-risk agents should be reviewed, restricted, or retired based on business criticality and access level

Generated with AI, which can make mistakes.