title: "Microsoft launches Project Perception, agentic security system" slug: "microsoft-launches-project-perception-agentic-security-system" published: "2026-08-14" beat: "Launches" tags: ["Launches"] creator: "Agentry Newsroom" editor: "Susanne Sperling, Editor — Human in the Loop" tools: ["Claude (Anthropic)", "Perplexity Sonar"] creativeWorkStatus: "verified" dateReviewed: "2026-08-14" aiActArticle50: "compliant" humanView: "https://agentry.news/launches/microsoft-launches-project-perception-agentic-security-system" agentView: "https://agentry.news/agent/microsoft-launches-project-perception-agentic-security-system"
Microsoft published Project Perception on July 27, 2026, introducing an agentic security system that coordinates red, blue, and green AI agents to investigate and remediate threats at machine speed. T
Drafted by an AI agent. Verified by Susanne Sperling, Editor — Human in the Loop. AI policy.
Microsoft published Project Perception on July 27, 2026, a new agentic security system designed for defending organizations in the AI era Microsoft. The platform enters public preview on August 3, 2026 Microsoft.
The system coordinates three types of AI agents working in continuous loops: red agents probe for weaknesses and vulnerabilities, blue agents investigate detected threats and anomalies, and green agents execute remediation or hardening actions SiliconAngle. Microsoft describes the architecture as "a continuously learning system of defense" that "brings together signals, context, models and specialized agents" while "keeping humans firmly in control" The Verge.
Project Perception is positioned as a departure from traditional alert-driven security tools. Rather than flagging events for human triage, the agents "reason, prioritize and act at machine speed," executing end-to-end security workflows autonomously Microsoft. The red-blue-green agent model allows simultaneous offensive and defensive simulation within a single platform: red agents identify attack paths, blue agents investigate those paths as if under real threat, and green agents automatically apply patches or configuration changes.
The public preview availability starting August 3 means organizations can begin testing Project Perception against their own security environments. Microsoft has not announced pricing or general availability timelines, but the preview phase signals the system is moving from research to production-grade engineering.
Project Perception is part of Microsoft's broader strategy to embed agentic AI across its security and cloud infrastructure. The announcement coincides with increased industry focus on autonomous threat response—a capability gap many enterprises have identified as critical as attack surfaces expand and response times compress.
For security teams, the implications are significant: agents that can act at scale and speed could reduce the mean time to remediate (MTTR) from hours to seconds. However, the approach also introduces new questions around agent alignment, audit trails for autonomous actions, and fallback procedures when agents conflict or make incorrect decisions—factors the preview period will likely expose.