AI systems are becoming the fastest data‑exfiltration vector inside the enterprise — and traditional DLP can’t see it.
This session is a live, end‑to‑end demo showing how to embed Microsoft Purview’s real‑time Data Loss Prevention directly into your AI agents and RAG pipelines.
Attendees will watch a three‑layer defense‑in‑depth architecture in action as it inspects prompts, retrieved documents, and model outputs — all in under two seconds. The demo highlights how delegated OBO authentication, smart redaction, keyword‑proximity scoring, and deterministic regex scanning work together to stop sensitive data leakage without breaking the user experience