ChatGPT CV and resume review that is worth having

The formatting test, which people skip

Do this before any content feedback, because it tests something the content feedback cannot.

Upload the exact file you send to employers and ask:

Read back the contents of this CV exactly as you received it: every heading, every job title, every date, in order.

If what comes back is garbled, sections are missing, or dates have detached from their jobs, that is a finding. Applicant tracking systems parse CVs the same broad way, by extracting text, and a CV that does not extract cleanly can be mangled before a human sees it.

The usual culprits are text inside columns, text boxes, tables used for layout, headers and footers, and text embedded in images. That last one is the worst: a name in a graphic is invisible to extraction entirely.

Prompts that produce real criticism

The rejection argument:

You are a hiring manager for [role]. Argue the case for rejecting this candidate at the CV screen. Be specific and cite the CV.

The ten minute screen:

You have sixty applications and ten minutes. What do you notice about this one in the first fifteen seconds, and would it make your shortlist? Why not?

The evidence check:

Which claims here are stated without evidence? For each, what would evidence look like?

This one is consistently the most useful. "Led a team" and "led a team of six through a migration that cut deploy time by half" are the same claim with and without evidence, and CVs are full of the first kind.

The gap check:

What questions would an interviewer ask about this CV that the candidate would rather not be asked?

Unemployment gaps, short stints, unexplained moves. Better to find them now.

Tailoring to a job

The strongest use of this, and it is a comparison rather than a review.

Here is a job advert and my CV. Produce a table: requirement, evidence in my CV, strength of that evidence, and where it appears. Mark any requirement I have not evidenced at all.

The unevidenced rows are the output. They are either things to add, because you have done them and did not say so, or things to be honest with yourself about.

Then:

Rewrite my three most relevant bullets to speak directly to the top three requirements, without inventing anything I have not claimed.

That last clause matters. Without it you get impressive bullets describing someone else.

What it cannot tell you

Whether you will get the job. It has no idea about the other applicants, the internal candidate, or the budget.

What is normal in your industry or country. CV conventions differ enormously. Length, photographs, personal details, what goes at the top. It will apply a general convention confidently.

Whether your experience is genuinely strong. It can tell you whether your CV communicates your experience. Those are different questions and only one of them is a document problem.

Keep the file clean

Since parsing is part of what is being tested, a plain, single-column layout in a standard font survives every system it meets. Elaborate design is worth much less than it looks, and it is what breaks extraction.

If ChatGPT reads your CV back perfectly, most automated systems will too, and that is worth more than any single piece of wording advice on this page.