Applying to mismatched job descriptions is a time sink. The average unoptimized resume is missing 52% of its target JD’s keywords (ResumeAdapter ATS data, 2026), and 99.7% of recruiters use those keyword filters to sort applicants. These 17 prompts help you score fit honestly, isolate the 2-3 deal-breaker skills, decide whether to apply, and tailor your resume plus talking points when you do.
TL;DR
- Paste the JD and your resume in one prompt so the model can cross-reference; two-shot prompting loses the links.
- Run the fit score + gap analysis + deal-breaker trio as a 5-minute triage before investing interview prep.
- Apply at roughly 6+/10 with no deal-breaker miss. A deal-breaker miss sinks even a 9/10 elsewhere.
- For tailoring honesty, Claude Opus 4.7 is the safest default (it does not fabricate metrics); ChatGPT GPT-5.5 is fastest for volume; Gemini 3.1 Pro is best when you can name the company and want live market context.
- Aim for a 65-75% keyword match to the JD, but never invent responsibilities to hit it.
What these prompts solve
A JD has 20-30 bullets but only 4-6 actually matter. Apply-to-everything wastes weeks; apply-only-where-you-fit-100% wastes opportunities. These templates separate must-haves from wishlist, score your honest fit, surface bridgeable gaps, and turn the analysis into customized resume bullets and interview talking points.
Which AI model to use (June 2026)
All three frontier models read a full JD plus a resume in one shot. The differences that matter for job-search work:
| Model | Plan & price (USD/mo) | Context | File upload | Best for |
|---|---|---|---|---|
| GPT-5.5 (ChatGPT) | Plus $20 | ~320 pages in-app (full 1M only on $200 Pro) | up to 512 MB/file | Fast bulk triage; aggressive action verbs |
| Claude Opus 4.7 / Sonnet 4.6 | Pro $20 ($17 annual) | 1M tokens | up to 30 MB/file | Honest tailoring; will not invent metrics |
| Gemini 3.1 Pro | Google AI Pro $19.99 | 1M tokens | up to 100 MB/file | Live company/salary research via Search |
Notes (as of June 2026): ChatGPT Free runs GPT-5.5 with tight limits and now shows ads in the US. Claude Pro bundles Claude Code and Cowork. Google AI Pro is the plan formerly called “Gemini Advanced” (renamed early 2026). A practical caveat: GPT-5.5 will sometimes fabricate a number such as “drove 40% revenue growth” when you did not supply one, so always fact-check any metric it adds. The honesty rule below is there for exactly this reason.
Who this is for
Job seekers triaging 30+ open roles a week, career switchers reality-checking whether their background bridges to a new industry, senior engineers comparing two offers’ JDs to decide which loop to invest in, new grads decoding inflated JDs (“5 years experience for an entry-level role”), and recruiters helping candidates prep tailored applications.
When not to use these prompts
Skip them when you already know the role is a clean match; the analysis won’t surprise you. Skip them when the JD is too vague to parse (a one-paragraph “we’re looking for great engineers”); ask the recruiter for the real expectations first. And don’t treat the fit score as a hard cutoff at 7+. Some 6/10 fits convert because of network, story, or hiring urgency, so judgment still matters.
Prompt anatomy / structure formula
A JD-matching prompt should always carry six elements:
- JD text: full, pasted, including the company section.
- Your background: resume or a condensed background dump.
- Goal: score, must-have extraction, gap analysis, or customization.
- Honesty rule: the model must call out gaps as gaps; “transferable skills” is not a free pass.
- Output shape: structured (table or numbered list), not prose.
- Action: what you’ll do with the result — apply, skip, prep, or customize.
Best for
- Pre-apply triage (apply or skip?)
- Cover-letter prep (what to emphasize)
- Interview-prep prioritization (which gaps to bridge first)
- Career-switch reality check (does your background bridge?)
- Comparing two roles head-to-head
- Decoding inflated or jargon-heavy JDs
- Spotting red flags hidden in JD language
- Tailoring resume bullets per application
17 copy-ready prompt templates
Placeholders in [brackets] are yours to fill. Paste the full JD and resume; the model needs both in one message.
1. JD-fit score
JD: [paste]. My resume: [paste]. Score the fit 1-10. Break down: 3 must-haves I clearly meet, 3 must-haves where I'm partial, 3 nice-to-haves I have, 3 gaps. End with a one-sentence recommendation: apply / apply with caveats / skip.2. Extract the actual must-haves
Below is a JD with 30+ bullets. Identify the actual must-haves (likely 4-6), separate from the wishlist. Mark each "must / strong-prefer / nice." Justify each "must" classification in one line.
JD: [paste]3. Apply or skip decision
My fit score is [N]/10 against this JD (paste). Should I apply? Give a recommendation with: (a) realistic chance of a recruiter screen, (b) what to emphasize in the cover letter, (c) what would change your mind if I provided one more piece of background.4. Gap-bridging plan
I have these 3 gaps vs the JD: [gaps]. For each, suggest 2-3 ways to bridge before the interview (e.g., a 1-week side project, a 1-hour course, a story I can tell honestly from adjacent experience). For each suggestion: time-to-execute and the credibility it earns.5. Talking points for the interview
Given the JD and my background, generate 8 talking points I should weave into interviews. For each: a one-line talking point, one specific example from my background (with a metric if possible), and which round to use it in (recruiter / hiring manager / panel).
JD: [paste]
Resume: [paste]6. Likely interview questions
Given the JD, predict 12 likely questions across: technical depth (4), behavioral (4), role-fit (4). Mark which are deal-breakers if answered weakly. Suggest one talking point per question I should prep.7. Compare 2 JDs to my background
I'm comparing JD-A vs JD-B (both pasted). Score my fit for each (1-10). Recommend which to prioritize and why. Include: which is more bridgeable if I'm under-fit, which has lower competition, and which optimizes for what I want next.
JD-A: [paste]
JD-B: [paste]
Resume: [paste]8. Translate JD-speak to plain English
This JD is dense and jargon-heavy. Translate it into plain English: what does the role actually do day-to-day? What does each "responsibility" look like in week 1 vs month 6? Highlight any phrase that hints at a real expectation buried in jargon.
JD: [paste]9. Identify hidden red flags
Below is a JD. Identify any phrases that hint at red flags: scope creep, low headcount for the work, vague seniority, unclear ownership, "wear many hats", odd compensation hints, "fast-paced" doublespeak. For each: the phrase, what it likely signals, and what I should ask in the recruiter screen.
JD: [paste]10. Tailor resume bullets for this JD
Given the JD and my resume, list specific bullet rewrites that would improve fit without lying. Show original then rewrite for each. Limit rewrites to bullets in my last 2 roles. Do not add fake responsibilities; only re-emphasize what was already there. Aim to surface JD keywords I genuinely match.11. Cover-letter outline tailored to the JD
Given the JD and my background, outline a 250-word cover letter: hook (1 line tying my background to the team's mission), 2 paragraphs of evidence (each citing one specific story or metric that maps to a JD must-have), close (1 line on why this team, not just any team).
JD: [paste]
Background: [paste]12. Spot the deal-breakers
From this JD, identify the 2-3 deal-breaker requirements — items where failing to meet means I won't pass screen no matter how strong the rest of my profile is. For each: the JD language, why it's likely a deal-breaker (regulatory, team-skill-gap, role-specific), and how to verify with the recruiter.
JD: [paste]13. Match probability and timeline
JD: [paste]. My background: [paste]. Estimate: probability my application gets a recruiter call (0-100%), probability I'd pass a phone screen, probability I'd get an onsite. Justify each estimate in one line. Be conservative; don't inflate.14. JD red-flag check vs Glassdoor
Below is a JD, then 8 Glassdoor reviews of the same company. Cross-check: which JD claims (e.g., "collaborative culture", "growth opportunity") contradict review themes? List contradictions with the review quote that contradicts each.
JD: [paste]
Reviews: [paste]15. Career-switch viability
I'm switching from [current field] to [target field]. JD: [paste]. My resume: [paste]. Honest assessment: which 2-3 skills will I be hardest pressed to demonstrate, and what's the minimum bridge each requires (project, certification, story)? Be honest; don't talk me into a stretch I'd fail.16. Salary expectation calibration vs JD
JD: [paste]. Geographic location: [location]. Seniority: [level]. Estimate the realistic salary band for this role based on (a) the JD's seniority signals, (b) the company's funding stage or market position if mentioned, (c) the location. Provide low / mid / high. Note the signals you used.17. JD-derived 30-day plan
JD: [paste]. Assume I get the role. Write a 30-day plan: week 1 (people to meet, docs to read), week 2 (first delivery), week 3 (first quick win), week 4 (first 90-day proposal). I'll bring this to the final round as a credibility-builder. Keep it specific to this team, not generic.Common mistakes
- Apply-to-everything strategy. Splatters effort, drops customization quality, and recruiters notice.
- Reading JD bullets as equal weight. 2-3 bullets are deal-breakers; the rest are wishlist. Don’t spend interview prep on the wishlist.
- Ignoring the 2-3 deal-breakers. Failing one means the application dies in screen; everything else is moot.
- Over-tailoring to the point of lying. Rewriting “wrote internal docs” as “led documentation strategy” gets caught in interviews. (Most ATS rejection is actually about missing keywords and broken formatting, not auto-reject rules: a 2025 study found 92% of recruiters set no content-based auto-reject, but tables and multi-column layouts cause about 23% of parse failures.)
- Cover letter that repeats the resume. It should add evidence that doesn’t fit in resume bullets, not restate them.
- Skipping company research. Two identical JDs can mean very different team cultures. Glassdoor plus LinkedIn employee tenure tells you more than the JD.
- Trusting the JD’s seniority label literally. “Senior” at startup A is often “Mid” at company B. Compare by responsibilities, not titles.
How to push results further
- Run the fit score, gap analysis, and deal-breaker prompts as a triage triple. Together they make the apply/skip call in under 5 minutes per JD.
- Always show the JD and the resume in one prompt. Two-shot prompting (analyze the JD, then compare to the resume) loses the cross-references the model needs.
- For roles you decide to apply to, run template #10 then #11 in sequence; the tailored bullets feed the cover letter.
- Use template #14 (Glassdoor cross-check) before final-round interviews. It surfaces what to probe in the “any questions for us?” slot.
- For career switches, run template #15 first. Most career-switch failures are predictable from a 5-minute viability check.
- After the recruiter screen, re-run the deal-breaker prompt with what you learned; the real deal-breakers often differ from what the JD claimed.
- Save your gap-bridging plan (template #4) as a permanent doc. Even if you don’t get this role, the gaps repeat across similar JDs.
FAQ
- What fit score should make me actually apply? Roughly 6+/10 with no deal-breaker miss. Below 6, the cover letter has to do unreasonable work. With a deal-breaker miss, even a 9/10 elsewhere won’t pass screen.
- Which AI model should I run these in? Claude Opus 4.7 (Pro, $20/mo) is the safest default because it won’t invent metrics. ChatGPT GPT-5.5 (Plus, $20/mo) is fastest for high-volume triage. Gemini 3.1 Pro (Google AI Pro, $19.99/mo) wins when you can name the company and want live salary and market context via Search. All three read a full JD plus resume in one message.
- Can I trust the model’s salary estimate? As a range, yes. As a precise number, no. Use it to set a floor, then verify with Levels.fyi or peer references.
- What keyword match rate should I target? A 65-75% match to the JD is the common benchmark, since the average unoptimized resume misses 52% of its target keywords. Hit it by re-emphasizing real experience, never by inventing responsibilities.
- Should I tailor every resume bullet per JD? Tailor only the top 2 roles’ bullets. Tailoring deeper produces over-customized resumes that look fake.
- Is the apply-or-skip decision really worth 10 minutes per JD? Yes for senior roles, where each application costs 4+ hours of interview prep. No for junior roles, where volume matters more than precision.
- What about ATS keyword optimization? Use resume keyword matching prompts for that; this article focuses on fit and customization, not ATS gaming.
Related
- Resume prompts — the broader resume rewriting workflow
- Resume keyword matching prompts — ATS keyword optimization layer
- Cover letter prompts — generate the cover letter from the JD-match analysis
- Cover letter customization prompts — adapt one cover letter across roles
- Company research prompts — pair with template #14 for deeper company diligence
- STAR interview prompts — story bank to back up the talking points from template #5
- JD matching analysis use case — end-to-end example
- Behavioral interview prep use case — full prep loop combining JD analysis and STAR
Tags: #Prompt #Job search #JD matching