Meridian Line's port-operations copilot logs full request and response payloads through AI Gateway for audit purposes, but some of those payloads contain a planner's personal contact details entered in free text. The compliance team wants the audit log retained while limiting who can see the raw PII inside it. Which combination of controls addresses this?

The requirement is to retain an audit trail while restricting access to sensitive contents. Keep AI Gateway payload logging enabled, then apply Unity Catalog access controls—and masking where supported—to the logged data so only authorized, need-to-know roles can view raw PII. This governs the audit log like other sensitive platform data rather than treating logging and governance as mutually exclusive. Disabling payload logging entirely may prevent PII from being captured, but it also removes the audit record the compliance team wants to retain. Storing logs outside Unity Catalog keeps the PII exposure problem while removing the access controls that could address it, along with the lineage and permission enforcement Meridian Line relies on elsewhere. Disabling payload logging and relying on rate limits and request counts also fails: rate limits cap request volume, and aggregate traffic records contain no payload content and cannot reconstruct planner activity as an audit trail. Exam caveat: masking is available only where supported, so it does not replace access controls on the logged data. Practical check: verify that logging remains enabled and test that an unauthorized role cannot view raw payload contents while an authorized role can. This pairing of AI Gateway logging with Unity Catalog governance preserves the audit trail while protecting PII.