How to Write a Cover Letter With ChatGPT (and Make It Sound Human)

A ChatGPT cover letter takes ten seconds to generate and about ten seconds for a recruiter to recognize. "I am thrilled to apply for this exciting opportunity" is not a sentence anyone writes on purpose, yet it opens thousands of letters a day. The tool is not the problem. The inputs and the editing are.

This guide shows you how to use ChatGPT for what it does well, which is structure, drafts, and tightening, while you supply what it cannot: real achievements, verified company knowledge, and your voice. You will get an input checklist, copy-and-paste prompts by paragraph, prompts for common situations, a humanizing edit pass, and a list of the red flags recruiters actually notice.

Plan on 30 to 45 minutes for your first letter and about 15 minutes for each tailored version after that.

Does a ChatGPT cover letter still work?#

Cover letters still get read, and tailoring still pays. In a ResumeGo field experiment covering 7,287 applications sent between 2019 and 2020, applications with a letter tailored to the job and company received 53% more interview callbacks than applications with no letter, and 87% of the hiring professionals it surveyed said they read cover letters, per ResumeGo's cover letter study. The same survey found most respondents valued tailored letters far above generic ones.

That is the core issue with an unedited ChatGPT draft. It is fluent but generic, and generic is exactly what the study found readers discount.

Why generic letters blend in more than ever#

Recruiters are reading more applications with fewer people. Greenhouse's benchmark data shows average applications per job more than doubled from 2022 to 2025, per its recruiting benchmarks.

What the research does and does not show#

The ResumeGo experiment used fictitious applications sent to real job postings, and it predates today's AI writing tools. It does not prove that any letter helps, and it cannot tell you how a specific employer weighs letters today. What it does show is directional: a letter written for the specific job and company outperformed sending nothing, and hiring professionals said generic letters hurt more than they help. That is the standard your ChatGPT draft has to meet.

Why AI-written letters face a higher bar#

AI writing is now an everyday tool. Pew Research Center found that 28% of employed US adults had used ChatGPT for work by early 2025, up from 8% in 2023, per its 2025 survey. With AI tools in wide use, a reviewer may see dozens of letters that open the same way. A specific first sentence and one concrete result are now what make a letter worth finishing. If you are deciding whether to write one at all, are cover letters necessary covers when they matter most.

What ChatGPT does well and badly on cover letters#

ChatGPT is a fast, fluent drafter with no knowledge of your career and unreliable knowledge of the employer. Split the work accordingly.

Treat it like a junior coworker#

OpenAI's own prompting guide compares a model to a junior coworker that does its best work when given explicit instructions for a specific output, per its prompt engineering guide. A junior colleague asked to "write a cover letter for me" would produce something vague too. Give it the material and the rules, and it produces something useful.

Gather your personalization inputs first#

Personalization is not a tone. It is information. Before you open ChatGPT, collect the facts that make the letter about you and this employer.

The inputs that matter#

This table lists what to gather and where each piece usually ends up in the letter.

You can pull the requirements quickly with the free job description keyword finder. For the greeting, how to address a cover letter covers what to do when you cannot find a name.

Where to find company facts that are actually useful#

• The company's newsroom or press page for launches, expansions, and funding
• The team's own blog or engineering, design, or product posts
• The careers page's description of the team and how it works
• Recent earnings materials or annual reports for public companies
• Conversations with current employees, which are the best source of all

Pick facts that connect to the role. "You recently opened a Denver warehouse" matters for a logistics analyst; it is trivia for a front-end developer.

If you cannot find useful company facts#

Small companies and stealth startups may have little public information. In that case, use what the posting itself reveals: the problem the team is hiring to solve, the tools it uses, or the customers it serves. "Your posting says the team is rebuilding onboarding for small-business customers" is a verified fact, because you read it. You can also mention how you use the product, if you do. Never ask ChatGPT to fill the gap with general claims about the company's culture or mission; those are the details most likely to be wrong.

An input sheet you can paste#

A step-by-step workflow to write a cover letter with ChatGPT#

Here is the full process from inputs to final file. Each step is short, and each produces something you can check.

1. Set the rules. Paste a session brief that bans invented facts and stock phrases.
2. Analyze the posting. Ask for the top requirements with supporting quotes from the job description.
3. Map your evidence. Ask which of your achievements best proves each requirement.
4. Draft paragraph by paragraph. Opening, proof, fit, close, with options for each.
5. Assemble and cut. Combine the best options into one letter under your word limit.
6. Run the humanizing edit. Replace generic language with your own.
7. Fact-check. Verify every claim about you and every fact about the company.
8. Format and send. Match the application method: PDF, text box, or email.

Reusing a letter for similar roles#

Once you have one strong letter, you do not need to start over for every application. Keep the proof paragraph as a reusable block, since your achievements do not change between employers. Rewrite the opening and the fit paragraph for each company, because those carry the personalization. Then run the claims check again; reused letters are where wrong company names slip in.

Why paragraph by paragraph beats one big prompt#

A single "write my cover letter" prompt returns a complete, smooth, forgettable letter. Drafting each paragraph separately, with three options each, gives you more control and more raw material. It also makes the humanizing edit easier, because you are choosing and combining rather than fixing a finished piece.

Copy-and-paste ChatGPT cover letter prompts#

Replace everything in [brackets]. Run the prompts in order within one chat dedicated to one job.

1. Session brief#

2. Requirements with quotes#

Asking for the quote keeps the analysis tied to the actual text. Anthropic's prompting guide recommends a similar technique, having the model pull relevant quotes before it does the task, per its prompting best practices.

3. Evidence mapping#

4. Opening paragraph options#

More approaches for first lines are in how to start a cover letter.

5. Proof paragraph#

6. Fit and closing paragraphs#

Closing options and sign-offs are covered in how to end a cover letter.

7. Assembly#

Prompts for specific situations#

Some letters need to explain something the resume cannot. These prompts handle the common cases. Pair each with the dedicated guide for the full approach.

Career change#

See career change cover letter for structure and examples.

Little or no experience#

The full guide is how to write a cover letter with no experience.

Internal role or referral#

For internal moves, cover letter for an internal position covers what to say about your current manager and team.

Employment gap or return to work#

A cover letter is a good place to address a gap briefly because you control the framing. Keep it to one or two sentences and return to evidence right away.

Short letter for a text box or email#

Short formats and when to use them are in short cover letter examples. If you are sending your application by email instead of a portal, how to write a job application email covers subject lines and structure.

The humanizing edit pass#

This is where a ChatGPT draft becomes your letter. Do it in passes rather than all at once, and expect to change at least a third of the words.

Pass 1: replace the opening#

Delete the first sentence the model wrote, even if it seems fine. Write your own opening in one or two sentences: what role, why you, why them. Specific beats clever.

Pass 2: swap adjectives for facts#

Search for words like "passionate," "dedicated," "proven," "exceptional," and "excited." Replace each with a fact that shows the quality. "Dedicated to customer satisfaction" becomes "kept a 96% CSAT across 1,200 tickets last year."

Pass 3: remove anything that fits every applicant#

Read each sentence and ask: could another candidate send this exact sentence? If yes, cut it or add a detail only you could write.

Pass 4: vary rhythm and read aloud#

AI drafts often use sentences of similar length with the same structure. Combine two short ones, split a long one, and start a sentence with something other than "I." Then read the letter aloud. Rewrite anything you would not say in a conversation.

Pass 5: check the length#

Cut until the letter fits on two-thirds of a page or the text box limit. Cutting is the most underrated edit; it removes filler and makes the remaining claims stronger.

Pass 6: add one detail only you know#

The most human sentence in a letter is usually one no model could have written: a feature of the product you rely on, a talk by the team lead you watched, a customer problem you have seen from the other side. Add one, briefly, where it supports your reason for applying.

Keep it true and keep it short. One genuine detail does more than a paragraph of praise.

AI tells and how to fix them#

Use this table to spot the patterns recruiters associate with unedited AI writing.

A before-and-after edit#

The edited version names the role, gives a reason tied to a verified company fact, and proves fit with a number, all in two sentences.

Red flags recruiters spot in AI cover letters#

Recruiters rarely run detection software; they notice patterns and mismatches. These are the ones that most often sink a letter.

• Wrong or invented company details. A product the company does not sell, an outdated mission statement, or a value the company never published.
• A letter that contradicts the resume. Different titles, dates, or numbers in the two documents.
• Perfect fluency with zero specifics. Polished sentences that never name a project, tool, or result.
• The job posting played back. Requirements restated as claims, in the same order, with the same phrasing.
• Placeholder residue. "[Company Name]," "Dear Hiring Manager at [Company]," or an unfilled [CHECK].
• Mismatched tone. Grand language for a casual startup, or casual language for a law firm.

For broader errors beyond AI use, see cover letter mistakes that hurt your application. For what reviewers notice in AI-assisted documents generally, read can employers tell if your resume was written by AI.

Follow-up prompts to refine the draft#

The first assembled letter is rarely the final one. Short, targeted follow-ups fix most problems faster than rewriting your original prompt.

If the letter looks like this, reply with this#

Ask it to critique, not just rewrite#

A useful final prompt is: "Read this as the hiring manager. What is the single strongest reason to interview me, and what is the biggest doubt this letter leaves?" If the answer to the first question is not what you want them to remember, move that point earlier. If the doubt is fixable with one more fact, add it.

Know when to stop prompting#

After three or four rounds, more prompting tends to polish rather than improve. If the letter still feels flat, the missing ingredient is information, not wording. Go back to your input sheet and add a sharper reason for applying or a better result.

Fact-check every claim about you and the employer#

A cover letter makes two kinds of claims: about you and about the company. ChatGPT can get both wrong.

Claims about you#

• Every number matches your resume and your records
• Titles, dates, and team sizes are exactly what your employer would confirm
• Verbs like "led" and "owned" describe your role, not your team's
• No tool, certification, or project appears that you did not provide

Claims about the company#

Language models can state outdated or invented company facts with full confidence. Treat every company detail in the draft as unverified until you have checked it on the company's own site or a recent, reputable news source. If you cannot verify a fact, delete it.

Watch for these errors in particular:

• A product, feature, or office the company does not have, or no longer has
• A mission statement or set of values paraphrased from memory rather than the company's site
• An executive or hiring manager's name that is wrong or out of date
• Awards, rankings, or growth figures with no source
• A competitor's product described as the employer's

A self-check prompt#

The prompt catches most drift. It does not replace your own check, because the model can misattribute sources too. The free cover letter checker gives a separate read on personalization, structure, and tone.

Privacy: what to share with ChatGPT#

To write a cover letter, ChatGPT needs your experience and the job, not your identity. Keep it that way.

What to remove before pasting#

• Your name, phone number, email, and street address
• References' names and contact details
• Salary history and current compensation
• Confidential details about your current employer, clients, or unreleased work
• Any government ID numbers or dates of birth

Check your data settings#

Most AI assistants let you control whether your chats can be used to improve their models. Anthropic, for example, says consumer chats are used for training only if you allow it through its Model Improvement setting, per its privacy center. ChatGPT and other tools have their own data controls and temporary or private chat options. Review them before pasting your resume.

Keep job searching off work accounts#

If your current employer provides an AI assistant, do not use it for your job search. Company tools may log activity, and drafting a letter to a competitor on a work account is a risk even where policy technically allows it. Use a personal account on a personal device.

Format, length, and where the letter goes#

The best letter still fails if it lands in the wrong format. Match the application method.

Choose the right format#

• Portal upload field: Save as PDF with a clear name, such as "Firstname-Lastname-Cover-Letter.pdf"
• Portal text box: Paste plain text with no greeting block; keep to the box's limit
• Email application: Put a short version in the email body and attach the full letter only if asked
• No field at all: Use the portal's "additional information" box, or send a brief note to the recruiter

Many portals offer a cover letter field or an open text box near the end of the form. Tips for each major system are in how to apply through Workday, Greenhouse, and Lever. Layout and spacing rules for a full letter are in cover letter format.

How long each part should be#

A strong letter spends most of its words on proof. The model below shows a typical word budget for a 300-word letter.

If your draft spends more words on enthusiasm than on evidence, rebalance it before sending.

A complete example, start to finish#

Here is a short letter assembled with the prompts above and edited with the humanizing pass. The details are fictional but realistic.

It runs about 190 words in the body, opens with a verified company detail, proves fit with numbers, and closes with a specific ask. For more complete examples across roles, see how to write a cover letter.

What changed from the first ChatGPT draft#

• The opening sentence was rewritten from scratch around a verified company change
• Three adjectives were replaced by account count, revenue, and churn figures
• A paragraph restating the posting's requirements was cut entirely
• The closing line became a direct question with a time frame

Your next steps#

Start with inputs, not prompts. For the next job you want, fill in the input sheet: three requirements, three achievements with numbers, and two company facts you verified yourself. That 15 minutes is what makes the letter specific.

Then run the prompts paragraph by paragraph, assemble the best options, and do the five-pass humanizing edit. Search for placeholders, check every claim, and match the format to the application method.

Reuse what works. Save your session brief and best prompts, and keep a short library of proof paragraphs you can adapt. The same evidence-first approach powers your resume too; see ChatGPT resume prompts for section-by-section prompts. If you would rather start from your existing resume, the free cover letter from resume tool drafts a matching letter you can then edit with the passes above.