Permission-aware retrieval
Answers only from documents the person asking is allowed to open.
Dropped: no access
Use whenPeople need answers from documents with different access rights.
AI where the work genuinely needs language or judgement. Nowhere else.
Only when it pays backYour documentation is spread across a file share, a wiki, a ticketing system and years of email. People know the answer exists; finding it costs more than rewriting it. A model can help, but only if it respects who is allowed to read what, and only if someone can say whether it works.
Answers only from documents the person asking is allowed to open.
Dropped: no access
Use whenPeople need answers from documents with different access rights.
The agent prepares the action. A person decides whether it runs.
If no: it stops
Use whenThe action touches money, customers, or records.
Agreed test cases decide when a system is ready, not a demo.
Fail: fix and run again
Use whenBefore anything reaches production.
The task follows fixed rules with no judgement in it: that is automation, cheaper and deterministic. Or your documents are out of date and contradict each other: retrieval will surface the wrong answer faster. Fix the source of truth first.
That is what the first call is for. Thirty minutes, no slides, and an honest answer even if it is not us.