How to Answer "How Do You Use AI?" in a Job Interview

Ellie at Willa, Editorial lead, Willa · Published September 2, 2026 · Updated September 3, 2026

Somewhere in your next interview loop, probably within the first thirty minutes, someone is going to ask how you use AI in your work. It has become a standard screen, the way "tell me about a time you handled conflict" was a decade ago. And most candidates answer it badly, either by rattling off tool names or by insisting that AI could never replace their judgment, which answers a question nobody asked.

Here is the good news: this question is unusually easy to prepare for, because interviewers are listening for something specific and most of your competition does not know what it is. This guide gives you a 3-part formula, sample answers for non-technical roles, and the traps that make strong candidates sound weak, so the AI question becomes the easiest part of your interview instead of the part you dread.

Why is every interviewer suddenly asking about AI?

Because employers quietly stopped treating AI skill as a bonus. In ZipRecruiter's 2026 AI Employer Report, 74% of employers said AI skills are now a strong advantage or an outright requirement for at least some roles, 13% require them for every position at the company, and half expect candidates to show up already being practical or advanced AI users.

The expectation has also spread well beyond technical jobs. LinkedIn's 2026 Skills on the Rise report puts it plainly: employers "are hiring for professionals who can deploy, manage, govern and commercialize" AI tools. Marketers, project managers, recruiters, and operations leads are getting this question just as often as engineers, sometimes more, because their roles are changing faster.

AI has even taken a seat on the other side of the table. In a Greenhouse survey of nearly 3,000 job seekers, 63% said they had already been through an interview conducted by AI itself. When the process is this saturated, "how do you use AI?" is not small talk. It is a real evaluation with a real rubric.

What is the interviewer actually listening for?

Not a list of tools. An interviewer who hears "I use ChatGPT, Claude, and Canva" learns almost nothing, because access to tools is universal now. What they are listening for is evidence of three things: that you actually use AI on real work, that it produced a result someone could measure, and that you supervise it instead of trusting it blindly.

There is a tension hiding inside the question, and knowing it is your advantage. In the same ZipRecruiter report, 57% of employers said AI has raised their baseline productivity expectations, yet only 22% provide mandatory AI training for all employees. They need people who taught themselves. At the same time, the human skills employers rated as more important than a year earlier were critical thinking (65%), judgment and decision making (59%), and creativity (58%). Read those two findings together and the rubric writes itself: they want someone who uses AI heavily and still thinks for herself. Your answer has to prove both at once.

How should you structure your answer? The 3-part formula

Keep it to about 45 seconds and hit these three beats in order.

1. Task. Name one specific, recurring piece of your real work where AI is now part of your process. Specific beats impressive: "first drafts of our customer onboarding emails" lands better than "content creation."

2. Result. Attach an outcome a manager would care about. Time saved, volume increased, quality improved, a decision made faster. Even an honest estimate ("what took me a morning now takes under an hour") works.

3. Judgment. Say how you check the output and name something you deliberately do not use AI for. This is the beat most candidates skip, and it is the one that separates "uses AI" from "can be trusted with AI."

As a fill-in-the-blank template: "In my current role I use [tool] for [specific task]. It cut [the work] from [X] to [Y], which freed me up to [higher-value work]. I treat it like a fast first draft, so I always [verification step], because it tends to get [specific weakness] wrong. And I keep [sensitive or high-stakes task] fully manual."

What does a strong answer sound like?

Here are three examples you can adapt, written for the roles Willa members actually hold.

A marketer: "I use Claude for campaign first drafts and for repurposing. One webinar becomes an email sequence, social posts, and a landing page outline in about an hour, where that used to take me two days. I never publish a draft untouched, because the voice drifts generic fast, so I rewrite against our style guide and fact-check every claim. Performance analysis I still do myself first, then use AI to pressure-test my read of the numbers."

An operations or project manager: "AI handles my meeting follow-ups. I run transcripts through it to pull decisions, owners, and deadlines, then post the summary within an hour of the call instead of the next day. I verify anything with a date or a number against the recording, because transcription errors compound. Vendor negotiations and performance conversations stay AI-free; those depend on context a model does not have."

Someone whose current employer restricts AI: "My company limits AI use with client data, which I follow strictly. So I have built my skills on my own work: I use ChatGPT to prep for stakeholder meetings by having it argue the other side's position, and I have taken what I learned into how I structure proposals. I would be glad to bring that habit somewhere that has clearer guardrails for using it fully."

Notice what all three do. They name one real task, attach a result, and volunteer a boundary without being asked. None of them oversell.

What mistakes make a good candidate sound bad?

These five patterns come up constantly, and each one is avoidable.

1. The tool list. Naming five apps with zero outcomes reads as "I have logins," not "I have skills."

2. The true believer. "I use AI for everything" makes an interviewer wonder what you contribute. Overselling AI undersells you.

3. The resister. "Honestly, I don't really trust it" positions you against the direction 74% of employers say they are heading. Skepticism about specific outputs is a strength; blanket resistance is a red flag.

4. The privacy blind spot. Casually mentioning that you paste customer data or confidential documents into public tools can end an otherwise great interview. If you handle sensitive data, say how you protect it.

5. The missing checker. Any answer without a verification step implies you ship whatever the model says. Always include how you catch its mistakes.

What if you barely use AI at work yet?

Then start this week, because one true story beats three invented ones, and you can build a true story in about two weeks. Here is the plan.

1. Pick one recurring task you mildly dread: status updates, meeting notes, first drafts, research summaries.
2. Do it with Claude or ChatGPT three times this week, adjusting your prompt each time.
3. Write down the honest before and after: minutes spent, drafts needed, what improved.
4. Note one error the AI made and how you caught it. That detail becomes your judgment beat.
5. Assemble the three beats (task, result, judgment) and say your answer out loud until it sounds like you.

If you want a broader on-ramp, our guide to AI skills for non-technical professionals maps out where to start, and once your story is solid, here is how to put those AI skills on your resume so the interview question never comes as a surprise. And if you would rather practice with real people first, Willa's live events are full of women rehearsing exactly these skills together.

Key takeaways

• "How do you use AI?" is now a standard interview screen: 74% of employers call AI skills an advantage or requirement, and half expect you to arrive already using them.
• Interviewers are scoring three things: real usage, measurable results, and judgment. Tool names alone score zero.
• Use the 3-part formula: one specific task, one concrete result, one verification habit plus a boundary.
• The judgment beat matters most, because employers rank critical thinking and decision making above almost everything else.
• No AI story yet? Build one in two weeks with a single recurring task and honest before-and-after notes.

The AI question is not a trap. It is one of the few interview questions where a small amount of deliberate practice puts you ahead of nearly everyone else in the pipeline. If you want company while you build these skills, join Willa. It is a community of women figuring out AI careers together, in public, at every level, and we would love to have you.

• job interview
• ai skills
• interview prep
• job search
• career development

Join Willa, the community where women learn AI together.

All posts