Agents Are the New Service Principals. Only More Dangerous | THR1803
Martina Grom presents a practical framework for AI agent governance, focusing on the security and compliance responsibilities that come with agents operating across systems.
Governing agents across Microsoft platforms
Organizations are deploying AI agents faster than they can govern them. Unlike traditional applications, agents can reason, take action, and interact across systems. Those capabilities make clear ownership, access boundaries, and operational oversight important governance concerns.
The session addresses governance across Microsoft 365, Entra, Purview, and Foundry, organized around four questions:
- Who owns the agent? Establish responsibility for each agent.
- What can it access? Define and control its access across systems.
- How is it monitored? Determine how its activity is observed.
- Who can shut it down? Identify who has the authority to stop it.
These questions provide a practical starting point for addressing agent security and compliance as deployment expands.
Session details and resources
- Event: Microsoft Ignite 2026
- Session: THR1803 — Agent Platform
- Format: Theater
- Level: 300 — Advanced
- Resource: Agent Platform next steps