How AI screens your CV — and how to write for it

AI is no longer the exception in hiring: many companies use AI-assisted evaluation to prioritise incoming applications against the job's criteria. On the candidate side this is mostly invisible — but knowing how it works helps you build a stronger application.

What these systems actually do

A typical AI-assisted evaluation compares the information in your CV with the requirements in the posting: do you have the skills asked for, does your experience length and field match, is the education requirement met, does the location work? The output is usually a match score with sub-scores; good systems also produce the reasoning behind every evaluation — which criteria are met and which are missing.

The most important thing to know: this is a ranking and prioritisation tool. In well-designed processes, the hiring decision — especially rejection — is made by a person, not the software. (At Kariora this is a design principle: the system scores, explains and suggests; rejecting a candidate never happens automatically.)

What this means for you in practice

1. The job posting is your best guide

Evaluation runs against the posting's criteria — not against some universal "ideal CV" template. Read the ad carefully: which skills are mandatory, how much experience, which languages? Does your CV show a visible counterpart for each?

2. Show skills in context

Modern systems do more than count words; they weigh how a skill was used within your experience. Don't just list "Python" in a skills box — say what you used it for, in which role, inside an experience bullet.

3. Keyword stuffing doesn't work — honesty pays twice

Sprinkling every word from the ad across your CV is an old trick that loses twice: an inconsistent CV produces weak reasoning in evaluation and collapses in the interview. Don't claim skills you lack; don't hide the ones you have.

4. Simple formatting is safe formatting

Evaluation works on your parsed, structured data. Complex tables, text boxes and text embedded in images can break parsing — a clean layout and a text-based PDF make sure your work fully reaches the system. Details in our ATS guide.

5. Treat screening questions as part of the application

The short questions at apply time (years of experience, work authorisation, certificates) are scored by rules in most systems. An empty or careless answer can outweigh a strong CV.

Should you be worried?

Healthy scepticism is fair; two things are worth separating.

Transparency: A black-box score and an explained evaluation are not the same thing. Where reasoning is produced, the hiring team can see why a score is what it is — the evaluation is auditable, and your application isn't compressed into a single opaque number.

Data rights: When you apply, you should see clear information about how your personal data will be processed, and your consent should be asked explicitly. Our guide to your data rights when applying for a job covers this from the candidate's side.

The takeaway

For a well-prepared candidate, AI-assisted screening is an accelerator, not a threat: an application that genuinely fits the posting and says so clearly stands out from the pile instead of disappearing into it. The rule hasn't changed — consistency just matters more: read the posting, describe your real experience clearly, send it in a clean format.