You didn't get rejected because you're unqualified. You got rejected because a piece of software scanned your CV for specific words, didn't find them, and never passed you to the person who would have known better. This is how almost every graduate application gets screened now, and most students are optimising for the wrong reader entirely.
An applicant tracking system, or ATS, is the software layer between you and the recruiter. It scans your CV for keywords lifted straight from the job description, scores the match, and ranks or filters candidates before a human opens a single file. Write a strong CV in the wrong words, and it never gets the chance to be read.
The scale of itQualified candidates, filtered anyway
This isn't a fringe problem. A widely cited 2021 Harvard Business School study, run with Accenture, surveyed employers directly about their own hiring software.
The system isn't malicious. It's literal. It's matching strings of text, not judging whether you can do the job — which means the fix is also literal: make sure your strings match.
The fixAsk AI to find the gap, not write the CV
The useful move here isn't asking AI to write your CV for you. It's asking AI to compare two documents you already have — the job posting and your current CV — and tell you exactly where the wording doesn't line up.
"Here's a job description and my CV. List every skill or keyword it uses that my CV doesn't mention, even if I have the experience worded differently."
You'll get a short, specific list back — not a rewrite, a gap report. From there, the edit is yours to make honestly.
Where this goes wrongThree things to not do with it
- Don't invent experience to match a keyword. The gap check tells you what's missing in your wording, not permission to claim something untrue.
- Don't let AI write the whole CV. A fully generated CV reads generic to the human who opens it after the ATS — and generic is exactly what gets you filtered at the next stage instead.
- Don't keyword-stuff. Repeating a term unnaturally can read as spam to the system and will definitely read as spam to the person reading it next.
Mirror the posting's actual language where it's true of you — "stakeholder management" instead of "worked with clients," if that's genuinely what you did — and leave the rest of the CV exactly as honest as it already is. The algorithm only needs to stop filtering you out. The human still decides.
Get the next one in your inbox
One AI tool, one prompt, three quick wins, every week. Free, and under 3 minutes to read.
Subscribe Free →