How do you handle a document changing or being deleted, and access being revoked?
Answer out loud or in your head first, then compare.
Answer
Each chunk should point to the document and version it came from. Then after a change I can rebuild the affected chunks, and after a deletion withdraw them from search and the cache. Updating the original file doesn't automatically remove all the copies.
A permission change may need a faster reaction than reprocessing the content. If access must be revoked almost immediately, a periodic import may not be enough: before using a result, I need an additional up-to-date check or another agreed blocking mechanism. Change notifications can get lost, so a periodic comparison of the source with the index is useful. I don't mix old content with accidental new metadata.
Why it matters: Shows that security changes over time.
Possible follow-ups: How do you find the derivatives of a deleted file? How do you know how big the update delay is?
How did it go? Be honest — “Not yet” puts the question on your review list.
Start with
- G-P01 · What is RAG? — RAG and retrieval
- G-P02 · What are embeddings? — RAG and retrieval
Further reading
- Azure AI Search: Document-level access control(opens in a new tab)Docs; accessed 22.10.2026. Security filters, native ACLs and preview features; syncing permissions needs an explicit mechanism.