ChatGPT prompts to tailor your resume to a job description

For software and IT roles · Updated August 29, 2026

Many job seekers ask how to use ChatGPT to tailor a resume to a job description. Paste-only prompts can speed up brainstorming — if you constrain the model to your facts. Used carelessly, ChatGPT will invent projects, metrics, and stack you never touched. This page gives copy-paste ChatGPT resume tailoring prompts for tech roles, explains what generic AI cannot see, and when a CV-grounded workflow like CVBranch is safer. For the full manual process, see how to tailor your resume to a job description; for match-first tooling, see match resume to job description.

Before you prompt: what ChatGPT does not know

A one-line prompt — "tailor my resume for this Software Engineer job" — fails because the model may not have your full PDF, cannot reliably separate required vs preferred skills, and has no way to verify a bullet against your story bank. Treat ChatGPT as a drafting assistant, not a source of truth.

Always paste your complete resume text (both pages if you have two) and the full job description into the chat. If the resume is a PDF, copy text or export to Word first — OCR paste is error-prone.

Ground rules for every prompt

Add this block to the end of any tailoring prompt:

Rules:
- Use ONLY facts from my resume and my answers below. Do not invent employers, dates, tools, or metrics.
- If the JD asks for something not on my resume, list it under "missing" — do not add it to bullets.
- Prefer one line per bullet: verb + scope + outcome.
- Mirror JD terminology only where my experience is genuinely the same.
- Flag any line you are uncertain about with [VERIFY].

Read every [VERIFY] line aloud as if you were in an interview. If you cannot defend it, delete it.

Prompt 1: Extract the JD (required vs preferred)

Run this first so later prompts target must-haves, not every buzzword in the posting.

You are helping me tailor a software/IT resume. Read this job description and output:

1. Job title and seniority signals (IC vs lead, years, scope)
2. Required skills/tools (must-haves from qualifications)
3. Preferred skills/tools (nice-to-haves)
4. Repeated themes in responsibilities (e.g. on-call, mentorship, distributed systems)
5. Top 5 must-have terms a recruiter might ctrl-F for

Job description:
"""
[PASTE FULL JD HERE]
"""

Save the output. You will paste the required/preferred lists into the match and rewrite prompts below. This mirrors what a resume–JD match report shows automatically.

Prompt 2: Match resume to job description

Manual overlap check before you rewrite anything. Ask for presentation gaps (hidden proof) vs real gaps (no evidence).

Compare my resume to the JD lists below.

For each required and preferred skill, classify:
- MATCHED (in skills or bullets with proof)
- PRESENTATION GAP (I likely have this but resume wording hides it — say why)
- REAL GAP (no evidence on resume)

Also note title/seniority fit and whether experience scope in bullets matches the JD themes.

Required skills:
[PASTE FROM PROMPT 1]

Preferred skills:
[PASTE FROM PROMPT 1]

My resume:
"""
[PASTE FULL RESUME TEXT]
"""

Rules:
- Do not suggest I add skills I do not have evidence for.
- Cite the resume line(s) you used for each MATCHED item.

Prompt 3: Rewrite one bullet (evidence-only)

Use per bullet, not on the whole resume at once — easier to verify. Pick a bullet that supports a JD theme from Prompt 1.

Rewrite this resume bullet for a [JOB TITLE] role. Target these JD themes: [THEME 1, THEME 2].

Original bullet:
"""
[PASTE ONE BULLET]
"""

Provide 2 variants. Each must:
- Stay one line if possible
- Keep the same facts, employer, and timeframe
- Not add tools, metrics, or scope I did not state
- Use JD terminology only where accurate

Rules: [PASTE GROUND RULES BLOCK]

Example (fictional): Original — "Supported production services at Example Co." JD theme — on-call, SLOs. A safe rewrite might mention incident response and latency only if those facts were in your gap answers — not because the JD asked for them.

Prompt 4: Clarify a presentation gap

When Prompt 2 flags a presentation gap (e.g. JD says .NET, resume says C#), ask follow-ups before drafting — same idea as a gap interview.

The JD requires [SKILL A]. My resume mentions [SKILL B] at [EMPLOYER] but not [SKILL A].

Ask me up to 3 short yes/no or factual follow-up questions to see if I have real [SKILL A] experience I forgot to list. Do not assume yes.

After I answer, either:
- Draft one bullet using ONLY my answers, or
- Say "real gap — do not add to resume" if evidence is insufficient.

My relevant resume lines:
"""
[PASTE BULLETS OR SKILLS SECTION]
"""

Rules: [PASTE GROUND RULES BLOCK]

Prompt 5: Order skills for this application

For stack-heavy JDs, reorder an existing skills line — do not add new tools.

Reorder my skills line for this JD. Use ONLY skills already on my resume. Put JD required matches first (in JD order), then preferred matches, then the rest. Max 10 skills.

JD required: [LIST]
JD preferred: [LIST]

My current skills line:
"""
[PASTE SKILLS]
"""

Rules: [PASTE GROUND RULES BLOCK]

Prompt 6: Trim to one page (selection, not invention)

Ask ChatGPT to select and shorten, not expand. Pair with your own length judgment.

I need a one-page resume version for this [JOB TITLE] application.

Given my full resume and the JD themes below, recommend:
1. Which 2 jobs to emphasize (most JD-relevant)
2. Max bullets per job (total ≤ 12)
3. Which bullets to drop or shorten (by exact text from my resume)
4. A 10-skill skills line (from my existing skills only)

Do not write new bullets — only select and trim from my text.

JD themes: [PASTE FROM PROMPT 1]

My resume:
"""
[PASTE FULL RESUME TEXT]
"""

Rules: [PASTE GROUND RULES BLOCK]

CVBranch Composer automates selection with JD-ranked bullets and a one-page fit hint — see Step 4 in the tailoring guide.

Prompts to avoid

• "Make me sound senior" — inflates title and scope without evidence.
• "Add keywords from this JD to my resume" — keyword stuffing; recruiters follow up in interviews.
• "Write me a resume for this job" — full fabrication risk.
• "Improve my metrics" — invented numbers are easy to catch.
• Upload-only without pasting text — you may not see what the model actually read from a PDF.

ChatGPT vs CVBranch (honest split)

ChatGPT is fine for: JD extraction, bullet wording variants, trimming prose you already trust, and interview prep from your own text. It is weak at: persistent match scores, required vs preferred parsing at scale, gap interviews tied to your story bank, and composing a one-page export with ATS preview — unless you rebuild that workflow in every chat thread.

CVBranch is built for software and IT CVs: upload once, paste the JD, see the match, answer gap questions, draft from verified facts, compose for that application. Matching is deterministic; AI assists drafting — you approve every change. Read can AI tailor my resume to a job posting? for a fuller comparison.

Quick checklist after ChatGPT output

• Every new skill appears in a bullet or gap answer you wrote.
• Metrics and scope match what you would say to a hiring manager.
• Required JD terms are covered with proof, not only in a skills list.
• No line relies on ChatGPT's assumptions — only on your paste and answers.
• You kept a master resume; this file is named for the company and role.