Microsoft unveils Guardian Agents to oversee enterprise AI agents

On October 19, 2026, Microsoft introduced a new security feature called Guardian Agents, designed to monitor the AI agents used by enterprises and guard against insider threats. The feature is part of Project Perception, a product line launched in June 2026 that initially included red team, blue team, and green team agents for simulating attacks, detecting vulnerabilities, and fixing them.

Microsoft CEO Satya Nadella has warned that enterprises should stop trusting AI model providers even when those providers offer safety assurances for their agentic products. He has urged companies to deploy separate systems, independent of AI labs, to watch and control the agents those products launch. Hayete Gallot, Microsoft's security business chief, said during the announcement that frontier models, regardless of their origin, cannot be trusted—not because they are malicious, but because they are unpredictable and often have access to sensitive data and tools.

Microsoft has also added agent detection to Microsoft Defender, which can identify up to 60 types of agents that employees bring into the workplace, including Claude Code, Codex, and OpenClaw. The company added features to limit agents' access to data and systems, and to dynamically adjust permissions based on the task, requiring human approval when an agent appears to be making risky decisions.

Microsoft is an inaugural member of Nvidia's Open Agent Safety Platform, which has the support of more than 100 tech companies. The platform isolates agents in secure environments and can monitor and stop behavior through hardware.