Microsoft Entra Ignite 2025 Session Replays for AI Security

Introduction: why this matters

As organizations accelerate AI adoption, identity becomes the control plane for trust—especially when AI agents begin acting autonomously across apps, data, and infrastructure. At Microsoft Ignite 2025, Microsoft Security and Microsoft Entra focused on how to extend Zero Trust beyond human identities to include AI agents and workloads, while improving detection and response for identity threats.

This post summarizes the key Entra replay topics and what IT pros should take away.

What’s new (and what to watch)

Secure access on the “AI frontier”

Microsoft’s Entra sessions emphasize that AI doesn’t require an identity “rebuild,” but it does require expanding Zero Trust controls to cover new identity types and access patterns:

• Microsoft Entra: What’s new in secure access on the AI frontier – updates on strengthening Zero Trust foundations and governing the growth of agents.
• Microsoft Entra Suite: Accelerate Zero Trust and secure AI access – identity-driven controls, adaptive policies, plus secure web and AI gateway capabilities to protect access to AI and agents.

AI agent identities: Entra Agent ID + Microsoft Agent 365

A major theme is managing agents like first-class identities:

• Secure access for AI agents with Microsoft Entra – focuses on discovering, managing, governing, and protecting agent identities using Microsoft Entra Agent ID.
• Microsoft notes that Entra Agent ID is included in the newly announced Microsoft Agent 365 (the “control plane for agents”) and works natively with Microsoft Foundry (formerly Azure AI Foundry).
• Additional sessions cover securing agents across their lifecycle (identity, posture, data access, governance) and practical, enterprise-scale implementation patterns.

Identity Threat Detection & Response (ITDR)

• Identity under siege: Modern ITDR from Microsoft – highlights tighter collaboration between Microsoft Entra and Microsoft Defender to protect the “identity fabric,” improve detection, and streamline response.

Security Copilot + Conditional Access optimization

• Security Copilot: Your agentic workforce for smarter, faster defense – includes a demo of a Conditional Access Optimization Agent in Microsoft Entra, pointing to more automated policy tuning and investigation workflows.

Integrated security foundation across Entra + Intune

• Top essentials for an Integrated, AI-ready security foundation – connects Entra and Intune scenarios to improve compliance and control across the Microsoft 365 stack.

Partner integrations (now GA)

Microsoft also called out General Availability of new Entra partner integrations (including Cloudflare, Akamai, Arkose Labs, HUMAN Security, and others), alongside partner sessions on deepfake/impersonation defense, NAC modernization, and cloud user environment management.

Impact for admins and security teams

• Expect more pressure to treat agent identities with the same rigor as users: lifecycle, least privilege, continuous evaluation, and governance.
• Conditional Access and identity posture will increasingly be evaluated through an AI-aware lens (where agents, apps, and automation are common).
• ITDR becomes critical as attackers target identities and helpdesk processes with AI-assisted social engineering.

Recommended next steps

1. Watch the Entra Agent ID / Agent 365 sessions and map where agents will exist in your tenant (creation, permissions, data access).
2. Review Conditional Access baselines and ensure policies cover workload/app identities and emerging agent scenarios.
3. Validate your ITDR readiness: alerting, investigation runbooks, and joint Entra/Defender workflows.
4. If relevant, evaluate newly GA partner integrations for phishing resistance, bot/impersonation defense, and modern network access controls.