Microsoft Unveils Project Perception, an Agentic Security System

Microsoft has unveiled Project Perception, an innovative security framework designed to counter AI-based attacks. Set to enter public preview on August 3, it features a multi-agent system with Red, Blue, and Green Teams that continuously learn and adapt to defend against threats.

Rather than relying on human analysis of security alerts, this architecture enhances response speed and reduces operational burden by incorporating a Security Context layer, effectively synthesizing threat intelligence and environmental awareness. This framework seeks to empower defenders by automatically implementing corrective actions while ensuring trust and compliance, particularly for regulated industries.

Microsoft aims to position itself competitively against other cybersecurity firms, leveraging its extensive data resources across various platforms. As an aspiring ecosystem rather than just a product suite, it intends to redefine enterprise security management.