From Risk to Resilience - Secure Your AI Agents with Microsoft Defender

Explore comprehensive security strategies for AI agents in this 42-minute conference talk from Microsoft Ignite 2025. Discover how AI agents are transforming modern workflows while creating new security and governance challenges that organizations must address. Learn about rising attack surfaces and novel threat vectors specific to agentic AI systems, and understand how these emerging technologies require specialized security approaches. Examine Microsoft Agent 365 as a centralized control plane for managing AI agents across your organization, and gain insights into attack path analysis using cloud resources and graph database technologies. Watch live demonstrations of three newly announced Microsoft Defender capabilities specifically designed to protect AI agents, including a detailed walkthrough of a Java customer helper agent's operations and security considerations. Observe real-time security scenarios including attacker simulation setups, responsible AI system responses that block attacks through Copilot Studio, and how Defender correlates security attempts to generate actionable alerts. Master best practices for confidently securing and governing AI agents in enterprise environments, with practical guidance from Microsoft security experts Ramya Chitrakar, Moran Gutman, and Asaf Nakash.