Using ChatGPT and AI Tools to Craft the Perfect Job Application
The prompts worth running in ChatGPT, Claude or Gemini for a resume, a cover letter and a final check, and where a model needs your facts, not its own.
You have the posting open and a resume that was written for a different job. ChatGPT, Claude and Gemini can close that gap faster than you can by hand, but only if you give them the right inputs and check what comes back. This guide gives you the prompts that are worth running for a resume, a cover letter and the final review, and the places where a model's answer needs a second look.
What These Tools Are Good At, and What They Are Not
ChatGPT, Claude and Gemini are the same kind of tool: a model that reads the text you paste and writes text back. For a job application, that makes them good at four things. They can read a posting and list what it asks for. They can turn a flat duty into a result-focused bullet when you give them the facts. They can draft a cover letter paragraph from your notes. And they can review two documents side by side and point out inconsistencies.
They are not good at knowing what you did. Every prompt below works only when you supply the specifics: the real numbers, the real tools, the real outcome. A model given "managed social media" will invent a result. A model given "grew the LinkedIn page from five thousand to twenty-five thousand followers in a year" will write a bullet you can defend in an interview. The rule for the whole process is that the facts come from you and the wording comes from the tool. Dedicated resume builders apply the same rule with the posting built in; the round-up of AI resume builders for 2026 compares seven of them.
Part 1: The Resume
Read the posting first
Paste the posting before you paste anything else, and ask for a structured breakdown. This is the input for everything that follows.
Read this job posting and list:
1. The hard skills and tools it requires
2. The soft skills and ways of working it emphasizes
3. The main responsibilities of the role
4. The exact phrases that appear more than once
5. Anything about the team or company that a cover letter could reference
Job posting:
[paste posting]
Keep the answer open in a second window. The repeated phrases are the ones a recruiter is likely to search for in the tracking system, and the ones your resume should use in the same words.
Rewrite one bullet at a time
Bullet rewriting is where these tools earn their place, and where they most often invent. Give the model the weak bullet, the facts behind it and the role you are aiming at, and ask for a rewrite that uses only those facts.
Rewrite this resume bullet so it leads with the result and names the
tools and method. Use only the facts I give you. Do not add numbers
I have not provided.
Current bullet: "Responsible for managing social media accounts"
Facts: ran LinkedIn and X for a B2B software company; grew LinkedIn
followers from 5,000 to 25,000 in twelve months; leads from social
went from about 20 a month to about 50 a month
Target role: Marketing Manager, B2B SaaS
An answer might come back as a bullet like "Grew the company LinkedIn audience from 5,000 to 25,000 in twelve months and more than doubled monthly inbound leads from social by building a data-led posting and engagement program." That is a bullet you can stand behind, because every number in it came from you. Run this for each of the five or six bullets that matter most for the posting, not for the whole document.
Compare the resume to the posting
Once the bullets are rewritten, ask for a gap review.
Compare my resume to this job posting. List:
1. Terms from the posting that my resume does not contain
2. Skills I clearly have, based on the resume, but have not named
3. Which bullets to move to the top of each role for this posting
4. Anything in the resume that could raise a question for this role
My resume:
[paste resume]
Job posting:
[paste posting]
A model's answer to this prompt is an opinion. For the plain count, the free resume keyword checker shows exactly which of the posting's terms your resume contains and which it is missing, in your browser, without sending your resume anywhere. Use the checker for the list of missing terms and the model for how to work the true ones in.
Write the summary last
The summary should be written after the bullets, because by then you know which terms the posting cares about and which of your bullets carry them.
Write a three-sentence professional summary for a resume targeting a
[job title] role at a [industry] company. Name the role, the years of
experience, and these three strengths: [skill 1], [skill 2], [skill 3].
Plain sentences, no adjectives like "passionate" or "results-driven".
My background: [two or three sentences about your experience]
Read the result aloud. If it sounds like a template, ask for a plainer version. If you would rather skip the prompting for the resume altogether, Resume Master does the rewrite from the posting in about two minutes and shows the keyword match before and after, so you can check every added term against what you actually did.
Part 2: The Cover Letter
A cover letter is read for two things: whether you understood the role and whether you can connect your experience to it. A model helps with structure and with the first draft. It cannot supply the connection unless you tell it what it is.
Research before you write
Based on this job posting and what I have pasted from the company's
website, identify:
1. What the team seems to be trying to accomplish this year
2. How this role contributes to it
3. Two facts about the company I could reference in a cover letter
Job posting: [paste posting]
Company pages: [paste the about page or a recent announcement]
Do not ask the model to tell you about the company from memory. It may describe an old version of the company, or a different company with a similar name. Paste the source text.
Draft the body paragraphs
The middle of the letter is the part that decides whether it gets read to the end. One paragraph per requirement, each built on a specific example.
Write one cover letter paragraph showing how my experience with
[specific skill or achievement] addresses [the posting's need].
Use this example, with its numbers, and keep it under 90 words.
What I did: [describe the work and the result]
Why it matters for this role: [one sentence]
Run it two or three times, once per requirement that matters most, then write the opening and closing yourself. A one-line opening that names the role and the reason you want it, and a closing that says you are available to talk, are better written by you than by a model that has written a thousand of them.
The authenticity pass
After the draft is assembled, run it back through the model once.
Revise this cover letter so it sounds like a person talking, not a
template. Remove clichés, remove any claim I did not make in my notes,
and cut anything that repeats the resume word for word.
[paste letter]
Then read it aloud. If a sentence is something you would never say in an interview, delete it.
Part 3: Prompts for Specific Situations
Changing careers
I am moving from [current field] to [target field] and applying for
[target job title]. Rewrite these bullets to emphasize the transferable
skills the posting names, using only what I actually did.
Bullets: [paste]
Posting terms: [paste the terms from Part 1]
The career-change sample resume shows what this looks like when the emphasis shifts and the facts stay the same.
Explaining a gap
I have a [duration] gap in my work history. During that time I
[what you actually did]. Write one sentence for my cover letter that
states this plainly and moves on.
One sentence. A paragraph draws attention; a sentence answers the question.
The final review
Before you submit, paste all three documents together.
Review my application for a [job title] role. Check for:
1. Dates, titles and company names that do not match between
the resume and the cover letter
2. Spelling and grammar
3. Terms from the posting that appear in neither document
4. Any claim in the cover letter that the resume does not support
5. Any bullet that describes a duty instead of a result
Resume: [paste]
Cover letter: [paste]
Job posting: [paste]
This prompt catches the errors a tired applicant sends: a company name left over from the last application, a date that changed between documents, a skill the letter promises that the resume never mentions.
Part 4: Tools Beyond ChatGPT
General-purpose models handle the writing. The other tools people use for applications fall into a few categories, and each one is worth understanding before you rely on it.
Keyword scanners compare your resume to a posting and report which terms match. They are useful for the count and no help with the judgment about which missing terms are true of you. The free keyword checker linked in Part 1 is one of these; it runs in the browser, needs no signup, and shows matched and missing terms.
Template builders give you a formatted document and a form to fill in. They solve layout, which matters, because tables, columns and text in headers break parsing in tracking systems. They do not tailor content, so you still need the prompts above or a tool that does the tailoring.
Auto-apply tools submit applications for you across many postings. They are the riskiest category. A resume that was not adjusted for the posting matches it partially, a form filled in by a script can carry the wrong answers, and a recruiter who receives a generic application and then a real one from the same person remembers the first. Sending fewer, better applications is the trade that pays off.
Resume Master combines the rewrite and the check: paste the posting and get your resume rewritten for it in about two minutes, with the keyword match shown before and after and updated live while you edit. It runs in the browser and on iOS and Android, and it never sends applications for you.
Mistakes That Waste the Tools
Pasting the output straight in
A model's first draft is a draft. Recruiters read many applications, and the ones written entirely by a model share a rhythm that stands out. Edit until it sounds like something you would say.
Stuffing terms
A recruiter's search looks for the posting's terms, and a person still reads the result. The resume examples show the right density: the posting's terms where they are true of you, inside sentences a person would write.
Letting the model invent numbers
If you did not give the model a number and a number appears in the output, delete it. Every figure on your resume will come up in an interview.
Ignoring formatting
ChatGPT, Claude and Gemini produce text. Put that text into a clean single-column document with conventional headings, or the parser on the other end will scramble it. The layout rules also change by country: a Canadian resume, for example, leaves off the photo and personal details and prints on letter-size paper.
Sending the same application everywhere
Faster tailoring is the point of these tools. It is not a reason to skip tailoring. Each posting gets its own pass through Part 1.
What to Do Next
Open the next posting you plan to apply to and run the first prompt on it. Rewrite the five bullets that matter most, using your own numbers, then run the gap review and the final check. That is one application done well, and the second one goes faster.
If you want the resume rewrite done in one step, paste the posting into Resume Master and read your resume's version of it in about two minutes. 3 AI credits free, no credit card needed, and every template and PDF export is free.
