AI Skills on Your CV and in Interviews: What Employers Actually Want

AI has arrived on CVs faster than employers have worked out how to read it. The result is a lot of applicants listing tool names under Skills and a lot of hiring managers ignoring that section entirely, because everyone has written the same thing.

What follows is what actually differentiates a candidate, and how to answer the questions now turning up in interviews for professional roles.

Why listing the tool does nothing

ChatGPT, Copilot and the rest take minutes to open. Listing one signals that you have used it, which is true of almost every applicant. It is the equivalent of listing email. A skills line reading AI, ChatGPT, Prompt Engineering tells a reader nothing they can act on, and in a regulated profession it can read as naive.

What differentiates is evidence of judgement: what you used it for, what you did not use it for, and what you did to make sure the output was right.

What employers are actually screening for

Three things, in roughly this order:

• Did you save real time on a real task? A specific, bounded outcome. Not transformed our workflow.
• Do you know where it fails? Candidates who can name the limits are treated very differently from candidates who are enthusiastic.
• Did you handle the confidentiality question? In finance, law, healthcare and HR this is often the first thing a hiring manager wants to hear, and it is the one most candidates never mention.

How to write it on the CV

Put it in the achievement bullets where the work is, not in a skills list. Compare:

Weak: Skills: ChatGPT, AI tools, prompt engineering.

Strong: Rebuilt the monthly variance commentary using an AI drafting step with a source-checking routine, cutting preparation from two days to half a day with no change to review findings.

The second version shows a task, a method, a measurable outcome and an implied control. It also survives the obvious follow-up question, which the first does not. Our guides to writing an accounting CV and creating a winning finance CV cover the surrounding structure.

One line of caution: do not claim a capability you cannot demonstrate in the room. AI claims are unusually easy to test, because the interviewer can simply ask you to walk through what you did.

Certificates: when they help

An AI certificate helps when the issuing body means something to the reader or when the course was assessed. It helps much less when it is a short self-awarded course, because the interviewer cannot tell what you had to do to get it. If you are weighing one up, our guide to AI certifications and our explainer on self-awarded versus accredited credentials set out what each kind is worth.

Be accurate about what a certificate is. Describing a half-day course as a qualification is the kind of overstatement that costs you credibility at exactly the wrong moment.

The interview questions now being asked

How do you use AI in your work? Answer with one specific task, the method, and the check you applied. Resist listing tools.

When would you not use it? The question they are really asking. Good answers: anything where I cannot verify the source; anything involving client-identifiable information in an external tool; anything where the judgement is the deliverable.

How do you know the output is right? Describe a process, not an instinct. Verifying facts and references at source, scaling the checking to the stakes — see our method for checking AI output.

Have you ever had AI get something wrong? Say yes and say what happened. A candidate who has never noticed an error has either not looked or not used it seriously. This is the question most often answered badly.

What is your view on AI replacing this role? They are testing composure and realism, not clairvoyance. The defensible answer is that it changes the mix of the work, moving effort from production toward review, judgement and accountability.

If you are early in your career

The anxiety that AI removes the junior rung is real and worth naming. The practical response is to be visibly good at the part that is not being automated: checking, questioning, and understanding why the work is done. A graduate who can explain why a number is wrong is more valuable than one who can produce the number faster, and that gap widens as production gets cheaper.

Frequently asked questions

Should I declare that I used AI to write my CV?

You do not need to declare it, but the CV must be accurate and sound like you. A generic application is the real risk, not the tool.

Is prompt engineering worth listing?

On its own, no. Shown as a method inside a concrete achievement, yes.

Do employers check AI claims?

Increasingly, by asking you to talk through the task. Claims that cannot survive that conversation do more harm than omitting them.

Does AI training count toward CPD?

Usually, if it is relevant to your role and you record it properly. Check your own body rules.

Our CPD courses cover AI skills in the context of professional work, which is the version that stands up in an interview.