For Job Seekers · Oct 4, 2026 · 2 min read
Can recruiters tell if AI wrote your CV? Yes, and here is what they notice
The tells are not the words. They are the shape of the claims, the missing specifics, and the mismatch with the interview. How to use AI for your CV without it reading like a machine wrote it.
Mário Rodrigues · IdealJob
Short answer: an experienced recruiter can usually tell, but not for the reason people think. They are not detecting the tool. They are detecting what the tool does when nobody checks its work.
The five tells
Adjectives instead of outcomes. "Dynamic, results-driven professional with a proven track record" says nothing a hiring manager can use. A person who did the work writes "cut the deploy time from forty minutes to six". Tools reach for adjectives because they don't know the number. You do.
The same shape in every bullet. Verb, object, vague benefit. "Developed scalable solutions to improve efficiency." Fifteen of those in a row read like a template because they are one.
Skills that don't match the experience. A list of twenty technologies and a work history that could only have touched six. Tools pad skills from the job posting. Recruiters cross-check, and the gap is the tell.
Nothing you could ask about. Good CVs invite questions: a project name, a client, a constraint. Generated ones are smooth and closed. There is nothing to grab.
The interview mismatch. This is the one that costs the job. The CV talks about leading a migration. The candidate, asked about it, describes watching one. From that moment every other line is doubted.
What a recruiter does not care about
Whether you used a tool to fix grammar, cut length, reorder sections or translate your CV into English. Whether the formatting came from a template. Whether a tool suggested a stronger verb. None of that is detectable and none of it matters. The CV is a document about your work; the tool is a word processor with opinions.
Using AI for your CV without the tells
- Write the facts first, yourself. Dates, roles, what you built, what changed because of it, with numbers wherever you have them. This is the part the tool can't do.
- Let the tool tighten, not invent. Ask it to shorten, to put the outcome first, to remove repetition. Reject anything it adds that you didn't give it.
- One version per role, from the same facts. Reordering for the posting is honest and effective. Rewriting your history for the posting is neither.
- Read it aloud. If a sentence is not something you would say in an interview, cut it or rewrite it until it is.
- Check the skills list against the experience. Every skill should be visible somewhere in a role.
How IdealJob's CV versions keep this honest
When the Career Agent suggests a CV version for a role, it works from the CV you wrote: it reorders skills, rewrites the summary around the role, and proposes changes to experience lines that you accept or reject one by one. It does not add experience, and it shows you the diff. The result is a CV a recruiter reads as yours, because it is, arranged so the relevant part comes first.
That is the whole trick. The tools that get candidates caught are the ones that write. The ones that help are the ones that edit.
Questions people ask
Do companies run AI detectors on CVs?
Rarely, and the detectors are unreliable. What companies do run is a person, who reads a hundred CVs a week and notices patterns.
Is it bad to use AI for my CV?
No. It is bad to let it invent, generalise, or write in a voice that isn't yours. Use it to tighten what you wrote, not to write what you didn't do.
What is the fastest way to make an AI written CV sound human?
Put a number, a name or a specific outcome in every bullet, and cut every adjective that you couldn't defend in an interview.