ChatGPT Prompts for Resumes

How to ask ChatGPT for help with a resume

«Improve my resume» gets you a resume that sounds better and says less. ChatGPT doesn't know which job you want, which results you can prove, or which lines you'd defend in an interview, so it guesses. It reaches for words like «spearheaded» and, if you're unlucky, a number nobody gave it. A resume request works when it hands over five things:

1. The job ad, in full. Requirements, duties, the company blurb. The employer's own words are what you're matching, and a summary loses them.
2. Your resume as plain text. Copy it out of your file so you both see the same words.
3. The numbers, from you. How many, how much, how fast, compared with what. If you don't know a figure, say so; don't let the model estimate it.
4. A rule against invention. «Use only the experience, titles, dates and tools in my resume.» Write it every time. It's the line that keeps a tailored resume honest.
5. What to hand back. The resume as plain text, then a list of what the ad asks for that you can't evidence. That second list is where your cover letter and interview prep start.

Put together, a request you can type from memory:

Tailor my resume to the job ad below.
Use only the experience, titles, dates, tools and numbers in my resume.
Where a bullet needs a number I haven't given, write [ADD NUMBER: what].
Put the evidence for the ad's top requirements in the top half of page one.
Return the resume as plain text, then a list of requirements I can't evidence.

Job ad: [paste]
My resume: [paste]

The prompts on this page are that request worked out for each job. The Job Ad Tailor is the pattern above in full; the Achievement Interviewer gets the numbers out of you before anything is written.

Which prompt for which job

One bullet, before and after

We wrote this example to show what the numbers do. Here's a typical duty line from a support job:

Responsible for handling customer support tickets and helping to improve response times.

It says what the job was and nothing about how well it was done. The Achievement Interviewer would ask how many tickets, what you changed, and what happened afterwards. In our example, the person answers: about 120 tickets a week, wrote 30 saved replies for the common questions, and first replies went from roughly nine hours to under three within two months. With those answers, the bullet becomes:

Cut first-reply time on a 120-ticket weekly queue from about 9 hours to under 3 in two months by writing 30 saved replies for the most common questions.

Every figure in it came from the person, and each one is something they can explain in an interview. Now picture the same line «improved» without that conversation: «Boosted customer satisfaction by 40% through process optimization.» It sounds stronger. It's also made up, and the first follow-up question in an interview would show it. If you don't have a number, the honest version is the bullet with [ADD NUMBER] left in until you find one, or a plain result without a figure.

ATS-friendly formatting, briefly

Applicant tracking systems parse your file into fields before a person sees it, and each employer's system behaves a bit differently. A few habits keep the parse clean almost anywhere:

• One column. Two-column layouts can come out with lines from both sides mixed together.
• Standard section names: Experience, Education, Skills. A heading like «Where I've made an impact» may not be recognized.
• Name and contact details in the body of the page, not in the header, and no text inside images or text boxes.
• One date format throughout, such as «Mar 2022 – Jun 2024».
• The ad's own terms, where they honestly describe your work. Once each, in a sentence, is enough.

The ATS Resume Scanner checks all of this from your plain text, and it's told never to suggest hidden text or keyword stuffing, which any person reading the resume would notice.

Cover letter and interview

The honest gaps list from the Job Ad Tailor is the best starting point for a cover letter, because it tells you what the employer will wonder about. A short letter that answers it beats a long one that repeats the resume:

Write a cover letter under 250 words for the job ad below.
Open with why this role, in one sentence, using a detail from the ad.
Use two results from my resume that match the ad's top requirements.
Address this gap honestly in one or two sentences: [gap].
Use only facts from my resume. Plain words, no "passionate".

Job ad: [paste]
My resume: [paste]

For the interview, take each requirement in the ad and find a story from your work that shows it. The STAR Answer Builder turns a rough story into a spoken answer and lists the follow-up questions you're likely to get. To practice out loud, the mock interviewer prompt asks you questions one at a time and waits for your answers. When the offer comes, the Salary Negotiation Script helps with what you say next.

Follow-up lines that fix a weak result

Where ChatGPT gets resumes wrong

The biggest risk with a resume is invention, and it rarely looks like lying. It looks like polish.

• Numbers appear. Asked for «quantified achievements», a model will quantify, whether or not you gave it figures. Read every number and ask: could I explain where this came from?
• Titles and scope grow. «Helped with» becomes «led», a team of two becomes «a cross-functional team». Tailoring to a senior ad makes this worse.
• Skills get borrowed from the ad. A tool listed in the job ad can slide into your skills section because it fits. If you haven't used it, it doesn't go in.
• Dates drift. Rarely, but a rewritten role can come back with a rounded or changed date. Check them against your original.
• Everything sounds the same. Polished bullets in an identical shape read like a template. Your own wording on one or two lines often helps more than a perfect verb.

A useful test for the finished document: could you talk for a minute about every line on it? If not, that line is the one an interviewer will pick.

Your resume and ChatGPT's memory

A resume holds your phone number, address and work history, so take the contact block out before you paste; the prompts don't need it. In ChatGPT you can also use a temporary chat. As of October 2026, OpenAI's help center says a temporary chat stays out of your history, doesn't create or update memories, and isn't used to improve its models while it stays temporary, though a copy may be kept for up to 30 days for safety purposes. You start one by opening a new chat and selecting Temporary. These prompts are plain text, so they also work in Claude, Gemini and other chat models; our ChatGPT review covers plans and limits.