Microsoft Launches Guardian Agents to Monitor Enterprise AI Agents

Microsoft has introduced security features including Guardian Agents, treating enterprise-deployed AI agents like insider threats for monitoring purposes. The feature is part of Project Perception, which Microsoft released in June 2026. Microsoft security chief Hayete Gallot said that frontier models cannot be trusted regardless of their origin, because they are trained to complete tasks rather than follow organizational norms. The core idea of Guardian Agents is to apply security monitoring and threat prevention to AI agents running within enterprises in the same way one would treat insiders. Microsoft has not yet disclosed pricing, a launch timeline, or early customers for the feature, and further developments remain to be seen.

AI overview of reporting · Updated 3 h前