Microsoft Project Perception Brings Coordinated AI Agents to Security Operations

Microsoft's Project Perception has entered public preview, bringing coordinated AI agents into security operations. These agents can investigate threats, assess risk, and connect findings with defensive actions.

The system is part of Microsoft's push towards coordinated security operations, combining security signals, organizational context, multiple AI models, and specialized agents in what the company calls a new Cyber Stack.

Project Perception consists of three classes of agents: Red-team agents identify possible paths to compromise, Blue-team agents investigate activity and determine meaningful risk, and Green-team agents take corrective actions to strengthen defenses.