Quick answer: paste the JD, add only facts you can prove, ask ChatGPT to map requirements to evidence, then draft and audit unsupported claims before you send.

Start with inputs, not a command

Give ChatGPT the job description, your resume notes, the company context you can verify, and the claims it must not invent. A bare write me a cover letter prompt usually produces a smooth but generic letter.

Split the prompt into four jobs

Use one pass to extract hiring signals, one pass to map evidence, one pass to draft, and one pass to audit. This keeps the model from hiding weak evidence inside polished paragraphs.

Force a send-before audit

Ask ChatGPT to list unsupported claims, invented motivation, keyword stuffing, and sentences that could fit any company. Fix those before you send.

Prompt

Act as a strict cover letter editor. Based on this JD and my resume notes, identify the 3 strongest hiring signals, map each signal to evidence I can prove, draft a concise cover letter, then list unsupported claims and generic sentences to remove. Do not invent metrics, company admiration, tools, seniority, or responsibilities.

FAQ

What should I paste into ChatGPT for a cover letter?

Paste the job description, 4 to 6 resume proof notes, the target role, and any real company reason you can verify. Do not paste private identifiers you do not need.

Can ChatGPT write the whole cover letter?

Yes, but only after it maps requirements to your evidence and runs an unsupported-claim audit. Otherwise the draft often sounds confident but interchangeable.