Angry customer reply checker prompt for any chatbot

The customer's message, your draft reply and your policy. In this prompt version you paste the text yourself. The steps are under "How to use it" below.

Works in ChatGPT, Claude, Gemini or any chatbot. Paste the customer's message, your draft reply and your policy (or what you are allowed to offer).

You are a careful second reader of a draft reply to an upset customer. Test the draft against the customer's message and my rules. Do not decide refunds. Never add a refund, credit, discount or timeline that I did not give you.

Customer message:
[paste]

My draft reply:
[paste]

Policy or what I am allowed to offer (write "none" if I have none, and then mark every promise "not checked against policy"):
[paste]

Case facts (optional): [dates, amounts, what happened]

Do this:
1. List each thing the customer asked for or complained about. Mark each Answered, Partly or Missed in my draft.
2. List every promise in my draft (refund, replacement, call back, time, "we will investigate"). For each, quote the policy that allows it, or mark it Not backed or Not checked.
3. Check the facts in my draft against the message and the case facts. Flag conflicts.
4. Flag tone problems, quoting my words: blame ("you should have", "as per our policy" used as a shield), defensive or corporate phrases ("unfortunately", "we regret any inconvenience"), an apology that names nothing specific, sarcasm, exclamation marks, jargon.
5. Check the draft says what happens next, who does it and by when.
6. Flag admissions of fault or liability for harm or legal matters, and say I should get advice if the customer mentions harm, legal action or a regulator. This is not legal advice.

Output: Verdict (Send as is, Send after fixes, or Hold and decide first), a table of what the customer raised and its status, numbered problems with my exact words and a fix, a repaired draft that keeps my facts and voice (mark open points [DECIDE: ...]), and a list of what needs my decision. Plain short sentences, no em dashes, no exclamation marks.

Before: Draft: "Unfortunately we regret any inconvenience. As per our policy, delivery times are estimates and not guaranteed, so you should have allowed more time. We will send you a full refund and a free replacement kettle within 24 hours."

After (skill output, from a real test run): Test run on a complaint and a draft we wrote, with a short policy. Verdict, unedited: "Hold and decide first. The draft promises a full refund and a free replacement, and the policy lets a support agent offer neither." It found the one thing the policy did allow, the 7.99 express refund, which the draft left out, quoted the blame phrases, and wrote a repaired reply with [DECIDE] markers for the refund timing and the team lead's approval.

• It checks the draft against the policy you give it. With no policy it marks promises "not checked".
• It does not decide refunds or know your approvals.
• It does not give legal advice. If a customer mentions harm, legal action or a regulator, get advice.
• It repairs wording. It will not make a bad outcome a good one.

Angry Customer Reply Check is also a skill you install once and call in plain words, so you do not paste the prompt each time. It can use its helper scripts and web access where your assistant allows. See the skill page, or read how to install skills.

See every skill for customer support teams.