An apparently inactive workflow can fail at several boundaries. A macro may have filled an unsaved draft, a trigger may have replaced the submitted field, or an automation may still be waiting for its hourly sweep. Trace the saved ticket and attributed events in that order. Repeating the same application without identifying the boundary can produce duplicate comments or confuse the audit. Repair the smallest incorrect step, then check the intended result.
Trace the evidence before changing a record. Each later task removes some guidance and introduces a different failure mechanism. Keep drafts, persisted fields, attributed events and derived results separate. A correct repair should explain why the expected outcome was absent and provide an observable check after the change.
Worked example: draft and event diagnosis.