By Peter Miller · Published September 1, 2026 · Last updated: September 20, 2026
10 min read
Ask a chatbot to "write me a resume" and you'll get something fluent, confident, and completely interchangeable with the two hundred other files in the pile. That's not because the tool is useless. It's because you asked it to invent a career instead of polish one you already have.
Knowing how to use ChatGPT and Claude to write a resume properly means feeding them bullets and a job description, then editing the output until every tool name and date survives. You bring the facts. The model brings phrasing. That split is the whole method, and it's faster than staring at a blank page.
Before you paste anything, check your resume for free and see what a parser already extracts from your current file. If titles and dates are scrambled, fix layout before you spend an hour on prettier bullets.
Below: what these tools can and cannot do, five ordered steps, copy-paste prompts, an edit pass that removes the generated tone, and mistakes that get files filtered before a human reads a word.
Quick Wins
- Before opening any AI tool, write ten real facts: tasks, numbers, tools, outcomes.
- Add "do not invent any numbers" to every prompt you use.
- Delete every adjective the model adds. All of them. See what survives.
What ChatGPT and Claude can do for your resume
They rewrite. They don't remember. A language model has no file on your career. It pattern-matches from millions of resumes that sound like yours. That makes it excellent at tightening a weak bullet you already wrote and terrible at knowing whether a claim is credible for your level.
I've screened stacks of AI-assisted files in Greenhouse, and the ones that pass read boring and specific: named tools, real dates, numbers you can defend in a phone screen. The ones that fail read like a press release nobody lived through.
| Good at | Bad at |
|---|---|
| Rewriting a weak bullet when you supply the facts | Knowing what you actually did last Tuesday |
| Spotting posting requirements you haven't addressed | Judging whether a claim fits your seniority |
| Translating internal jargon into market language | Producing specifics without inventing them |
| Cutting a long bullet to 20 words | Formatting a file that parses cleanly in Workday |
Content and formatting are separate problems. A model can give you excellent words inside a layout that fails at the first gate. For the layout half, see formatting mistakes that kill AI resumes . This page is the words.
Five categories you need per job before you prompt anything: scope (what you owned), outcome (what changed), numbers (real ones), tools (named), and context (industry, company size, what made it hard). Missing any one category is why the model fills the gap with fiction.
How to use ChatGPT and Claude to write a resume: five steps
The first step happens offline. Everything else is paste, edit, verify. If you only have ten minutes tonight, do steps 1 and 5. Facts first, parser check last.
Step 1: Collect raw facts before you open a chat
For each role, write what you owned, what changed, the number attached, and the tools involved. Ugly notes in a phone memo are fine. This is the material the whole process runs on, and no prompt replaces it.
Before: "Help me write a resume for a marketing manager role."
After: A note with ten lines: launched email A/B test, 14% lift in trial signups, Mailchimp and GA4, owned list of 42k contacts, reported to VP weekly.
Step 2: Rewrite bullets with hard constraints
One job at a time. Paste your raw notes, not the whole PDF. Ask for outcome-led bullets under 25 words, with an explicit ban on invented figures and a request to ask you for missing detail instead of guessing.
Before: Ran innovative cross-functional initiatives driving operational efficiencies.
After: Led invoice approval redesign across finance, ops, and IT; cut approval time from 9 days to 3.
Step 3: Tailor against one posting
Paste the job description and your experience section. Ask which requirements you address clearly, weakly, or not at all. Add "do not rewrite anything yet" so the model diagnoses gaps instead of inventing experience to close them.
Fill real gaps yourself. If the posting wants HubSpot and you used it twice, write that bullet with a date. If you never touched HubSpot, you don't add HubSpot. The gap analysis tells you where to spend your editing time, not where to lie.
Step 4: Edit out the generated tone
Generated resume text has a texture recruiters spot fast: grammatically perfect, relentlessly positive, and empty. Run four passes on anything a model produces.
- Delete every adjective. Dynamic, innovative, cutting-edge. Read what's left.
- Restore specifics. Put tool names, numbers, and the actual system back in.
- Swap throat-clearing verbs for plain ones. Led, built, cut, ran, fixed.
- Read aloud: could you talk about this line for two minutes? If not, delete it.
Before: Used data-driven insights to enhance stakeholder decision-making.
After: Built the weekly margin dashboard in Power BI that the commercial team runs pricing calls from.
Step 5: Check parsing and coverage
Words are half the job. Upload the finished file to a parser check and confirm titles, employers, and dates stored in order. Then compare coverage against the posting. A polished paragraph in a two-column template still dies silently if Workday reads your skills list as your employer name.
Copy-paste block: bullet rewrite prompt
You are an experienced recruiter. Rewrite the following as three
resume bullet options. Rules: lead with the outcome, under 25 words
each, name the tools I mention, use plain language, no adjectives like
dynamic or innovative, and do not invent any numbers or facts I have
not given you. If something important is missing, ask me for it
instead of guessing.
Here are my raw notes:
[paste your facts here]
Job description for context (do not invent experience to match it):
[paste posting here]
For cover letters, same rules apply with higher risk because generic letters are even more obvious. The cover letter generator builds one against a specific posting if you'd rather not manage the prompting yourself.
Career changers get real value here: ask the model to map your responsibilities onto the target field's vocabulary, then verify each translation is honest. A teacher's differentiated instruction across 30 students really is stakeholder management, but you have to defend the framing in an interview.
What not to paste: contact details, named clients under NDA, confidential figures, colleagues' performance data, or anything from a work account without checking employer policy. Anonymised facts produce the same quality of rewrite because the model works on sentence structure, not client identity.
Capture wins as you go with a weekly accomplishment log . The same notes become resume bullets whether you're applying or negotiating pay.
Where AI resume drafts go wrong
"Write me a resume for a marketing manager." You'll get the average of every marketing resume ever written, which is precisely what you're competing against. The model has no access to your Q3 campaign or your actual Mailchimp list size.
Accepting invented numbers. Models produce plausible figures when asked for impact. Check every single one against reality. If you didn't measure it, cut the number or replace it with scale: 6 sites, 40 invoices, 3 retainers.
Skipping the edit pass. The adjectives are the tell. Removing them takes four minutes and changes how the file reads in a stack of twenty.
Pasting the whole resume at once. You lose control of which sections get rewritten, you expose contact info and NDA client names, and the model averages everything into the same mid-level register. Paste bullets and the JD. Not the file.
Letting it choose the format. It will suggest a layout that doesn't parse. Keep your single-column template. Change words inside it.
Using the same generated file for every application. The tailoring step is the one that actually raises your match score. One master file, edited per posting.
Claiming skills to close a gap. Closing gaps on paper creates them in the interview. If the gap analysis says you're weak on Salesforce, write the truth: used it twice on a migration, or don't claim it.
Check the draft before you send it
A model can't tell you whether your file parses, and that's the gate that rejects most applications silently. Upload the finished draft to the free ATS resume checker and read the extraction. Every job present? Dates attached? Skills where they should be? Those three answers tell you whether the writing even reached the stage where it matters.
Then check coverage with the job match score against the posting you're targeting. Fix honest gaps in bullets, not in a keyword footer. Two tools, one workflow: parse first, match second.
Write ten facts first
- How to use ChatGPT and Claude to write a resume: you bring facts, the model brings phrasing, you verify everything.
- Paste bullets and the JD. Not your whole file. Not your contact block.
- Ban invented numbers in every prompt. Then delete every adjective. What survives is your resume.
Do this today: write ten real facts about your current role before you open any tool. That list is worth more than any prompt on this page.
Then confirm the draft works. Check your resume for free and see what the parser makes of it.
Read more
- How to beat AI resume screeners : the gates your draft has to clear.
- ChatGPT resume mistakes : failures to avoid in more detail.
- Formatting mistakes that kill AI resumes : the half no model can fix for you.
Frequently asked questions
Often, though not from detection software. They notice it from the writing: fluent, confident, and completely non-specific. A resume full of impact and cross-functional collaboration with no numbers, tools, or scale reads as generated whether or not it was.
Yes, as a drafting and editing assistant. You supply the facts, it improves the phrasing, and you verify every claim. What is not acceptable is inventing experience, and it fails immediately at interview because you cannot discuss work you did not do.
One that gives the model raw facts and hard constraints: the real task, the real outcome, the tools used, plus instructions to keep it under 25 words, lead with the outcome, and never invent numbers. Vague prompts produce vague bullets.
Strip your contact details, and remove client names, confidential figures, or anything covered by an NDA first. Paste the experience content you want rewritten rather than the whole document, and check your employer's policy if you are using a work account.
Passing depends far more on formatting and specificity than on who wrote the words. Generated text that lacks named tools, numbers, and the posting's actual terminology tends to score poorly, because there is nothing concrete for a matcher to latch onto.