Brand governance
Every operator answering customers in public is writing brand copy. Almost none of it is approved by anyone.
Corporate defines the voice in one place: the phrasing you want, the language you never want, and real example replies the model matches for rhythm and warmth. Any of it can be locked, so a franchisee cannot override it at their location.
Banned language is enforced twice — once when a reply is generated, and again immediately before it posts. So a phrase you add to the list this morning also catches drafts that were already written and waiting.
One arrangement worth knowing: every reply can wait for the operator's approval while the draft they are approving is locked by corporate. The words are always yours. The decision is always theirs.
Total control at corporate, full autonomy for your franchisees, or anywhere in between — set per setting rather than as one blunt switch. Every override request and every corporate action is logged, so the franchise relationship stays on the record instead of in someone's inbox.
When a location goes quiet, corporate can answer on its behalf, and the operator sees that it came from corporate rather than finding a reply they did not write.
Most platforms treat brand voice as a tone dropdown. That produces replies that are inoffensive and completely interchangeable — the same three sentences every business in the category is posting, which customers have learned to read as nobody being home.