Your next act
6 min read · July 31, 2026
Last verified: July 31, 2026
Resume vs. the Screener
Run your resume against a real job post and see what the ATS and the recruiter actually see, in about 90 seconds of AI time.
What this is for
You're applying to a specific job and want to know, before you hit submit, whether your resume clears the keyword filter and whether the recruiter skimming it for six seconds will see what they need to see. This is for that moment — not a general resume tune-up.
Before you start
- The full job posting, copied as plain text (not a link — paste the actual text)
- Your current resume, either pasted as text or uploaded as a PDF/Word file
- Your last three roles' bullet points, ready to paste separately if you're using Prompt 2 on its own
The prompts
1. Score it like the ATS and the recruiter would
Use either ChatGPT or Claude — this is straightforward text comparison, no advantage either way. Upload your resume file (in ChatGPT: click the paperclip in the chat box; in Claude: click the paperclip or drag the file into the chat) and paste the job post text into the same message.
ROLE: You are a corporate recruiter and an ATS keyword-matching system,
evaluated together.
CONTEXT: My resume is attached. Here is the job posting I'm applying to:
[PASTE FULL JOB POST TEXT]
CONSTRAINTS:
- Score match on a 0-100 scale the way an ATS keyword parser would score it
(literal term matches: skills, tools, certifications, job titles).
- Separately, score how a human recruiter would react in a 6-second skim
(clarity of seniority, relevance of top bullets, red flags).
- Identify exactly 3 keywords or phrases from the job post that are missing
or under-represented in my resume. Do not suggest more than 3 — I need
the highest-leverage gaps, not a wish list.
- Do not rewrite my resume in this response. Just score and diagnose.
OUTPUT FORMAT:
1. ATS score (0-100) with the 5 terms that matched and the 3 that didn't
2. Recruiter skim reaction (2-3 sentences, blunt)
3. The 3 missing keywords, each with the exact sentence in my resume where
it could honestly be added (no fabricating experience I don't have)
2. Rewrite your last 3 bullets as outcomes with numbers
Use either. Do this after Prompt 1 so the AI knows which keywords to work in.
ROLE: You are a resume editor who specializes in converting task lists
into outcome statements.
CONTEXT: Here are my last 3 bullets, verbatim:
[PASTE 3 BULLETS, e.g. "Managed a team that handled customer escalations"]
The job post I'm targeting emphasizes: [PASTE 2-3 PRIORITIES FROM POST,
e.g. "reducing churn, cross-functional leadership, budget ownership"]
CONSTRAINTS:
- Rewrite each bullet in the format: [Action verb] + [what you did] +
[measurable result]. If I didn't give you a number, ask me for one
instead of inventing it.
- Keep each bullet under 25 words.
- Do not add skills, tools, or scope I did not mention.
OUTPUT FORMAT: The 3 rewritten bullets, plus a flagged list of any bullet
where you need a real number from me before it's finished.
3. Compress 25 years into 2 pages without erasing seniority
Use either. This is the one people get wrong by either cutting too much (looks junior) or cutting too little (looks unfocused). Paste your full resume, even if it's currently 3-4 pages.
ROLE: You are a resume editor for senior candidates with 20+ years of
experience, targeting a 2-page format.
CONTEXT: Full resume pasted below. I am applying for [JOB TITLE] roles
similar to this one: [PASTE JOB POST TEXT]
CONSTRAINTS:
- Full bullet detail (3-4 bullets each) only for the 3 most recent or most
relevant roles.
- Roles older than 15 years get one line each: title, company, years —
no bullets, unless a role contains the single most impressive
credential of my career, in which case give it one bullet.
- Cut or merge redundant roles (e.g. two similar titles at the same
company) into one line.
- Do not remove graduation years unless I explicitly ask — removing them
selectively looks worse than leaving them, since gaps get noticed.
- Total output must fit 2 pages at 11pt, which is roughly 900-1,100 words
including headers.
OUTPUT FORMAT: The full compressed resume, followed by a one-line note
on what you cut and why.
4. Generate the 5 questions this job post implies they'll screen for
Use either. Run this last — it tells you what to rehearse before the call, not what to fix on the page.
ROLE: You are the hiring manager who wrote this job posting.
CONTEXT: Job post: [PASTE FULL JOB POST TEXT]
CONSTRAINTS:
- Infer the 5 questions you would actually ask in a first-round screen,
based on what the post emphasizes (not generic interview questions).
- For each, name which line or requirement in the post it's testing for.
- Order them from most likely to least likely to come up first.
OUTPUT FORMAT: A numbered list of 5 questions, each followed by one line
naming the requirement it tests.
The "dates back to 1998" problem
If your resume shows work history stretching back 20+ years, the fix isn't hiding the start date — recruiters know how math works, and an obvious gap or vague date range reads as evasive, not younger. The fix is asymmetric detail: give full bullets to the last 10-15 years, and reduce anything older to a single line with title, company, and years, no bullets. One exception — if an early-career role contains a credential that still matters (you built the system half the industry now uses, you were first in a role that didn't exist yet), keep one bullet for it. Prompt 3 above does this automatically; just don't let the AI talk you into deleting the line entirely.
A warning on keyword stuffing: adding the 3 missing keywords from Prompt 1 means working them into real sentences about real work, not pasting a skills list at the bottom in white text or cramming synonyms into a bullet until it reads like a Boolean search. ATS systems and recruiters have both gotten better at catching this, and a recruiter who catches it will read your whole resume more skeptically. If a keyword doesn't honestly fit anywhere in your actual experience, leave it out — that's a real gap, not a formatting problem.
Where this goes wrong
- You paste the job title instead of the full post. The AI needs the actual language of the posting to find real keyword gaps — "Senior Marketing Manager" tells it nothing. Always paste the full text.
- You let the AI invent numbers for Prompt 2. If you don't have a metric, say "I don't have an exact number, estimate a reasonable range and mark it as an estimate" or leave it as a scope statement instead. A fabricated number that comes up in an interview is worse than no number.
- You apply the same compressed resume to every job. Prompt 3's cuts are specific to the job post you fed it. Re-run it per posting, or at minimum per role type — a resume compressed for a "Director of Operations" post won't emphasize the right 3 roles for a "VP of Supply Chain" post.
The 2-minute version
Paste your resume and the job post into one message. Run Prompt 1 for the score and the 3 gaps. If the score's low, run Prompt 3 for the 2-page compression. Skim the Prompt 4 questions on your way to the interview.
Where you stand with AI, in 10 minutes. The free AI Readiness Quiz gives you a personalized roadmap instead of another tool list: genxcelerate.com
Weekly, on Tuesdays. The Gen Xcelerator — practical AI intelligence for people who already know how to do the job. Subscribe free
Experience Is the API. — GenXcelerate
Related resources
- Pricing Your ExpertiseSet a number you can say out loud without flinching, and know why it's right — 20 minutes of prep before your next quote.
- The Ageism-Proof Interview PrepFive prompts to prep for the questions that are really about your age, without ever saying the word "age" — done in one sitting the night before.
- The Fractional One-PagerTurn 25 years of experience into a document that can close a $5,000 engagement — draft it in 30 minutes, tighten it in 10.
The Gen Xcelerator
A new one every Tuesday. Short, useful, and written for people who already know how to run things.
Get the newsletter