Got an invite to an AI interview and wondering what it will ask? This page lists the AI interview questions and answers that come up most, from the opening "tell me about yourself" to the follow-ups a live AI interviewer asks after you speak. It's written for candidates preparing for the round and for hiring managers deciding what their AI screen should cover. Each question explains what good looks like, and the most common ones include a sample answer to adapt in your own words. Read the section on adaptive follow-ups first. It explains why the second question matters more than the first, and why a memorized script tends to fall apart.
What a strong AI interview checks
- Relevant experience: whether past work actually matches what the job needs.
- Depth under follow-up: whether the candidate can explain the details behind a claim.
- Structured stories: clear STAR answers with the candidate's own actions and results.
- Problem solving: reasoning out loud through a new problem, not just recalling facts.
- Communication: clear, direct answers at a steady pace.
- Coachability: how the candidate responds to a hint or a harder version of a question.
How adaptive follow-ups work
AI interviews come in two broad types. In a one-way format, the system shows fixed prompts and records your answers, and nothing changes based on what you say. In a live, conversational format, the AI interviewer listens to each answer and decides what to ask next. Our guide to what an AI interview is covers the difference in more detail.
In a live format, follow-ups usually do one of three things. They dig for specifics ("What was your part in that?"), they raise or lower difficulty based on how you're doing, or they test the edges of a claim ("What would break if traffic doubled?"). A vague answer tends to earn a narrowing question. A strong one tends to earn a harder one. Some tools, including NTRVSTA, also offer hints, so getting stuck isn't the end of the question: how you use the hint becomes part of what's measured. For how answers turn into scores, see how AI interviews are scored.
AI interview questions: introduction and background
1. Tell me about yourself and what brings you to this role.
What good looks like: a focused 60 to 90 second summary that ends on why this role fits. The AI will often pick up a detail from this answer and ask about it, so only mention what you're ready to discuss.
Sample answer: "I'm a backend developer with three years of Python, mostly building internal APIs for a logistics company. My biggest project was rebuilding our shipment tracking service so it could handle peak season. I'm applying because this role is closer to customer-facing products, and I want my work to reach users directly."
2. Walk me through your current role. What does a typical week look like?
What good looks like: concrete tasks, tools and who they work with, rather than a job title restated. Numbers help, such as team size or tickets handled.
3. What do you know about this position?
What good looks like: they've read the posting and can name two or three duties in their own words, plus a question they'd want answered later.
4. Why are you looking for a new role now?
What good looks like: brief and forward-looking, with no complaints about a past employer.
Experience and role fit
5. The role requires [key skill]. Describe a recent time you used it.
What good looks like: a real, recent example with enough detail to survive "tell me more." Strong candidates say how deep their experience goes instead of overselling.
Sample answer (for a data role asking about SQL): "Last month I wrote the queries behind our churn dashboard. The tricky part was customers with several accounts, so I grouped by billing ID instead of account ID and checked the totals against finance's numbers before we shipped it."
6. What's a tool or skill on your resume you'd rate yourself lower on?
What good looks like: honesty, and a plan to close the gap. Live AI interviewers often probe every skill listed, so this question rewards a resume that's accurate.
7. Tell me about the project you know best.
What good looks like: the goal, constraints, their decisions and the result, explained clearly enough for someone outside the team.
8. What kind of work environment helps you do your best work?
What good looks like: specific and honest (clear goals, quick feedback, focus time), and compatible with how the role actually works.
Behavioral questions an AI interviewer asks
Many AI interviewers (NTRVSTA included) structure these around the STAR method, so name the situation, your task, your actions and the result. More practice prompts are on our behavioral interview questions page.
9. Tell me about a time you faced a difficult problem at work.
What good looks like: a real problem with stakes, the candidate's own steps, and a result.
Sample answer: "Our support queue doubled after a pricing change, and response times slipped past two days (situation). As team lead, I had to get us back under 24 hours without new hires (task). I tagged a week of tickets, found 40 percent were the same three billing questions, and wrote help articles plus saved replies for them (action). Within two weeks we were back under a day, and those three topics dropped by half (result)."
10. Describe a time you disagreed with your manager.
What good looks like: they raised the concern respectfully, with evidence, and committed to the final decision even if it wasn't theirs.
11. Tell me about a goal you missed.
What good looks like: ownership, what they learned, and what they now do differently. Expect a follow-up like "What would you do in the first week next time?"
12. Describe a time you had to explain something complex to a non-expert.
What good looks like: they adjusted the language, checked understanding, and the other person could act on it.
13. Tell me about a time you helped a teammate succeed.
What good looks like: specific help (pairing, reviewing, covering work) and a result for the teammate or the team.
Technical and skills follow-ups
For technical roles, an AI interviewer may pull from a question library and then adjust. These are typical shapes, not a fixed list.
14. Explain [concept you mentioned] as if to a new teammate.
What good looks like: an accurate, plain-language explanation with an example. The goal is to check understanding, not vocabulary.
15. How would you debug [a slow page or a failing job]?
What good looks like: a calm, ordered process. Reproduce it, check logs and metrics, form a hypothesis, test one change at a time, and confirm the fix.
16. What trade-off did you weigh in that design, and what did you give up?
What good looks like: a real alternative they considered, why they rejected it, and what it cost them. "There were no trade-offs" is a weak answer.
17. Sketch how you'd design [a simple system] on the whiteboard.
What good looks like: requirements first, then the main parts and how data flows between them, then scaling and failure. Talking through each step matters as much as the diagram.
Adaptive follow-up questions
These are the follow-ups that catch candidates off guard. They aren't trick questions. They test whether the first answer was real.
18. Can you put a number on that result?
What good looks like: an honest figure or estimate ("about 30 percent fewer tickets, based on the monthly report"), or a clear statement that it wasn't measured and how they'd know it worked.
19. What was your part, specifically, versus the team's?
What good looks like: a crisp split. "The team shipped it. I wrote the migration plan and ran the rollback test."
Red flag: everything was "we," with no clear personal contribution.
20. What would you do differently if you did it again?
What good looks like: a real change based on what they learned, not "nothing."
21. Now suppose [the data were 100 times larger, or the deadline were cut in half]. What changes?
What good looks like: they adjust the plan instead of repeating it, and say what they would cut or add and why.
22. Here's a hint: [a nudge toward the approach]. How does that change your answer?
What good looks like: they take the hint, build on it, and explain why it helps. Using guidance well is a strength, not a penalty in itself.
A simple scoring rubric
Running this interview with AI
For hiring teams, NTRVSTA runs this kind of first round as a live, real-time AI interview, with a video avatar or audio only. It adapts questions and difficulty to each answer, uses STAR for behavioral questions, and offers hints so you see how candidates take guidance. Its exam library covers languages like Java, Python, C++ and JavaScript, frameworks and tools, and soft skills, and system-design interviews use an interactive whiteboard. You can customize questions per role and run interviews in multiple languages.
The report includes a score breakdown (technical skills, English proficiency, communication) and video snippets of key moments, so a recruiter can check the reasoning behind a score before deciding. Compare options in our AI interview software guide. Candidates preparing for any format can read how to prepare for an AI interview and our video interview questions.
Questions
Can I ask an AI interviewer to repeat or rephrase a question?
In a live conversational AI interview, usually yes. Asking for clarification is normal and shows you want to answer the right question. In a one-way recorded format you may not be able to, so read each prompt carefully and use your thinking time.
Is it okay to say I don't know in an AI interview?
Yes. An honest "I haven't used that, but here's how I'd approach it" is better than guessing. An adaptive interviewer will often adjust the next question, and your reasoning can still earn credit even when you don't know the answer.
Will an AI interviewer catch me using ChatGPT during the interview?
Don't test it. Using a hidden AI assistant is usually against the rules and can get your application rejected. Live follow-up questions also make it hard, because the next question asks about the details of what you just said, and you'll be expected to explain your answers again in later rounds.
Want to hear adaptive follow-ups before your real interview? Start a free sample AI interview. Running interviews for your team? Book a 20-minute demo with us.
