Updated:

Short answer: Quick answer: paste the JD, role title, resume notes, and one honest reason for applying. Ask ChatGPT for a signal-to-proof table first, then a short draft, then a final audit for generic claims, invented company praise, keyword stuffing, and lines you cannot defend.

Best for

Applicants who want a reusable ChatGPT prompt that turns one JD into a signal-to-proof map, a short draft, and a send-before audit.

Avoid if

Avoid using it as a one-click letter writer or to hide missing proof, invented referrals, fake metrics, or company research you did not verify.

What to do next

Paste the JD, role title, proof notes, tone, and reason for applying; run the table first, then draft and audit.

Start with a real application packet

Paste the JD, exact role title, resume notes, target tone, and one honest reason for applying. Ask ChatGPT for a brief before prose: hiring signals, proof available, proof missing, and the risk if it invents each missing claim.

Ask for a signal-to-proof table first

For each must-have requirement, pair one resume, project, portfolio, or work-habit proof. If proof is weak, the prompt should ask you a question instead of filling the gap with polished language.

Set the fact boundary and sign off every claim

Before drafting, separate verified facts, writing constraints, and private or forbidden details. Require [VERIFY: missing fact] wherever the evidence has a gap. After drafting, compare every employer name, role title, date, metric, tool, credential, referral, and company-specific statement with the JD, resume, or source note. Mark each one Keep, Rewrite, or Delete yourself before sending.

Prompt

Act as a strict cover letter editor. Use only the information inside <JOB_DESCRIPTION> and <EVIDENCE_NOTES>; obey <WRITING_CONSTRAINTS>. First extract 3 to 5 hiring signals. Build a table with signal, exact evidence, evidence source, confidence, and missing question. Ask up to 3 questions before drafting. If a fact is still missing, write [VERIFY: what is missing] instead of guessing. Then draft the letter within the requested length and tone. Finish with a claim ledger listing every name, date, metric, tool, credential, referral, and company-specific statement, its source, and my action: Keep, Rewrite, or Delete. Do not mark the letter ready to send until I confirm every ledger row.

FAQ

What is the best ChatGPT prompt for a cover letter?

The best prompt makes ChatGPT read the JD, map hiring signals to evidence, ask questions when proof is missing, draft briefly, and audit unsupported claims before sending.

Can I reuse this prompt for every job?

Reuse the workflow, not the content. Replace the JD, proof notes, role reason, tone, and send-before audit for each application.

Related pages