Microsoft Uses AI-Powered Agentic Scanning to Find Windows Security Flaws and Accelerate Patching

Microsoft is expanding AI-driven vulnerability discovery across Windows with its Multi-Model Agentic Scanning Harness (MDASH), combining multiple AI models and a two-stage pipeline (scanning and validation) to find and de-duplicate potential flaws earlier in the development lifecycle. The initiative integrates AI into the Secure Development Lifecycle, validation pipelines, and engineering workflows to accelerate patch creation and deployment while retaining human oversight; it also highlights customer-focused patch management tools and practices such as Windows Autopatch, hotpatching, Intune, Defender integrations, and Known Issue Rollback to reduce exposure windows.