Secure Govern and Scale: Protecting AI Agents with Microsoft Agent 365 | THR857
Brendan Powers presents an intermediate-level session on bringing agent context and security evidence together in Microsoft Agent 365. The demonstrations connect registered-agent security with Microsoft Defender investigation and partner perspectives on data and behavioral signals.
Agent security workflow
The session follows a practical workflow for securing AI agents:
- Discover: Identify registered agents and bring their context into the security review.
- Assess: Examine security evidence, including data and behavioral signals, to inform agent-risk assessment.
- Investigate: Use Microsoft Defender to investigate agent security.
- Protect: Choose supported responses based on the assessment and investigation.
Partner signals and proposed integrations
Cyera and Darktrace contribute perspectives on data and behavioral signals. The session explores how proposed partner integrations could inform agent-risk assessment; these are presented as proposals, not as confirmed available capabilities.
The response-selection discussion focuses on supported actions, keeping the distinction between potential integrations and available responses clear.
Session details and resources
- Event: Microsoft Ignite 2026
- Session: THR857 | Copilot
- Format and level: Theater | (200) Intermediate
- Copilot next-step resources