Ask a security question in plain English. The agent translates your intent into KQL, queries Microsoft Defender telemetry, and analyzes the results for potential threats.
This is a public demo — the AI query planning and threat-hunt analysis below are real (and run on the owner's OpenAI balance), against a bundled sample dataset rather than live production telemetry. It's still password-protected to prevent abuse. Browse the History page to see past findings without one.