A free prompt for anyone running an AI assistant that talks to upset customers. Give it the customer's message, the drafted reply and the next steps you allow, and it rewrites the reply: the customer's actual problem in their own words, one concrete action with the exact link or email you listed, and none of the stock empathy lines. It never blames the customer, never admits fault and never promises anything you didn't list. Threats, chargebacks and abuse are handed to a person. Written to match my prompt packs.