Unlocking Data Security with Microsoft Purview
summary
A sample customer‑service app uses the MIP File SDK to fetch labels, check view and extract rights, compute label inheritance by taking the max sensitivity value, then pass only authorized text to Azure OpenAI for summaries. Restricted chats are redacted and summaries inherit the highest label so unauthorized users see “sorry, I can’t summarize” instead of secrets. Purview AI Hub receives lightweight audit events for later eDiscovery and compliance.
AI-written summary · about
chapters (17)
- 0:00Microsoft Purview and AI security
- 1:17Data security in AI apps
- 3:10Microsoft security and safety stack
- 4:42Purview’s three pillars
- 6:04Copilot for M365 with Purview
- 7:11Protecting AI apps with labels
- 11:33Sample app demo overview
- 13:19Labels in application database
- 17:16Code deep dive on labels
- 21:11Extract rights in AI summaries
- 25:05Inherited sensitivity labels
- 28:22AI summaries with labels
- 30:33Copilot for your apps
- 31:39Purview AI Hub for auditing
- 33:13AI Hub integration demo
- 36:34Key takeaways
- 38:18Q&A on labels and auth
related talks
17mExtend data security into multi-cloud apps and solutions
09mProtecting AI Agents with Microsoft Purview
10mData Security & AI Adoption: A Leader’s Perspective (Customer Spotlight)
50mNotes from the field Microsoft Information Protection
23mSecure Microsoft 365 Copilot and agents against data loss
48mData Loss Prevention and Azure Information Protection...
49mThe Power BI secret sauce: Security and governance
12mStriking the Balance: Safeguarding Customer Privacy While Empowering Employees