A job interview is one of the highest-stakes conversations most people have. The outcome can change your income, your daily life, your sense of what's possible. And yet the average amount of structured preparation most people do is: read the job description once, Google the company briefly, and think about what to wear.
I've been on both sides of interview tables over the years — as the candidate, as the hiring manager, and as someone who coached friends through the process. The people who do well are almost always the ones who did their homework. Not because they were smarter, but because they walked in knowing things the other candidates didn't. They asked better questions. They connected their experience to what the company actually needed. They didn't get caught flat-footed by the obvious questions.
AI makes that level of preparation available to anyone willing to spend an hour with it the night before. Here's exactly how I'd approach it.
Why interview prep is where AI earns its keep
Most AI use cases involve some trade-off between speed and quality — you get something decent faster, but you'd get something better if you did it yourself. Interview prep is one of the few areas where AI can actually beat what most people would do on their own, because the task is research-heavy, the information is publicly available, and the output is a structured set of talking points rather than original creative work.
You're not asking the AI to be you. You're asking it to help you build a foundation so that when you walk in, you're spending your mental energy on the conversation — not scrambling to remember what the company does or nervously wondering what questions might come up.
That's a real thing. I've seen people tank interviews not because they were unqualified, but because they were visibly unprepared. And I've seen less-experienced candidates shine because they showed genuine curiosity about the role and the business. AI can help you get there.
Step 1 — Research the company
Start here, and give this more time than you think it deserves. You want to walk in understanding not just what the company does, but what they care about, where they're headed, and what pressures they're operating under. That context shapes every answer you give and every question you ask.
Open ChatGPT, Claude, or Gemini and start with something like this:
"I have a job interview at [Company Name] next week. Can you give me a summary of what they do, who their customers are, what their business model is, and any notable recent news or developments? I want to understand the company well enough to have an intelligent conversation about their business."
Then follow up with:
"What are the biggest challenges or competitive pressures a company like [Company Name] is likely facing right now? What would someone in [role title] need to understand about those challenges?"
And one more:
"Based on the job description below, what does this company seem to be trying to accomplish by hiring for this role? What problem are they probably trying to solve?"
[Paste the job description]
That last one is underrated. The job description tells you what skills they want. The question above helps you infer what's actually going on — why this role exists now, what success looks like, what's probably not working without it. That kind of insight is what makes a candidate sound thoughtful rather than just qualified.
One important note: AI works from public information. If the company isn't well-known or has minimal online presence, the responses will be thinner. Treat everything as a starting point and verify what matters. Don't walk in and confidently repeat something the AI told you without checking it first.
Step 2 — Anticipate the questions
Most interviewers work from a relatively predictable set of questions. Not because they're unimaginative, but because the classics exist for a reason — they surface useful information efficiently. "Tell me about yourself." "What's your biggest weakness?" "Describe a time you handled a difficult situation at work." You've heard them. The problem is most people treat them as hurdles to get past rather than opportunities to land specific points.
Use AI to generate a likely question list before you walk in:
"I'm interviewing for [role title] at [Company Name]. Based on this job description, what questions is the interviewer most likely to ask? Include both standard behavioral questions and role-specific questions that relate to the responsibilities listed."
[Paste job description]
You'll get a solid list — probably 15 to 20 questions. Some will be obvious. A few will be ones you hadn't thought to prepare for. Go through the list and flag the ones that make you slightly nervous. Those are the ones to focus on.
Then for each answer you want to prepare, you can use AI as a thought partner:
"Here's my draft answer to 'Tell me about yourself' for this interview. Can you give me honest feedback on whether it's focused, whether it connects to what the company seems to need, and whether anything should be cut or added?"
[Paste your draft]
The AI won't know your actual experience — you'll need to supply that. But it's good at pointing out when an answer meanders, when it's missing the so-what, or when a different structure would be cleaner.
Step 3 — Practice your answers
Reading through prepared answers is not the same as being able to say them out loud under mild pressure. The gap between those two things is where most people fall down.
One option is to use AI as a mock interviewer:
"I'd like to do a mock interview for a [role title] position. Ask me one question at a time, wait for my answer, then give me brief feedback before moving to the next question. Start with 'Tell me about yourself.'"
Type your answers as if you're saying them. It forces you to actually formulate a response rather than just knowing roughly what you'd say. When the AI gives feedback, pay attention — it will often catch things like vague answers, missing outcomes, or answers that don't address the question actually asked.
Even better: use a voice-capable AI tool for this. ChatGPT's voice mode on mobile, or any browser-based voice interface, lets you speak your answers aloud and hear them back. This matters more than people expect. An answer that reads well on paper can sound stilted or rambling when you actually say it. Speaking it out loud — even just to your phone — surfaces those rough edges before the interview does.
I'd recommend doing at least one full pass of your three or four most important answers out loud. Not in your head. Out loud. It's uncomfortable the first time. That's the point.
Step 4 — Prepare smart questions to ask them
"Do you have any questions for us?" is not a formality. It's an evaluation. Candidates who ask nothing, or who ask only about salary and vacation policy, signal that they're not particularly interested in the role itself. Candidates who ask sharp, specific questions about the business signal that they've thought carefully about whether this is the right opportunity — which, counterintuitively, makes them more attractive.
Use the research you've already done and ask AI to help you develop questions:
"Based on this job description and what you know about [Company Name], what are three to five genuinely interesting questions I could ask the interviewer? I want questions that show I've thought about the role and the business, not generic questions I could ask any company."
Then filter those through your own judgment. Pick two or three that you're actually curious about and that feel natural coming from you. Don't bring a printed list of eight questions and work through it methodically — that reads as prepared in the wrong way. Two thoughtful questions asked genuinely are worth more than ten rehearsed ones.
Some categories that tend to work well: questions about what success looks like in the first six months, questions about challenges the team is currently working through, questions about how this role connects to broader company goals. What to avoid: questions where the answer is obviously on their website, or questions that are really just thinly veiled negotiating moves.
What AI can't do
All of this is genuinely useful, and I mean that. But I want to be straight about the limits.
AI can't read the room. It can't tell you that your interviewer seems skeptical and you need to shift your approach, or that they lit up when you mentioned a particular project and you should lean into that. The actual interview is a live human conversation, and no amount of prep replaces the ability to listen, adapt, and respond to what's actually in front of you.
AI also can't manufacture genuine enthusiasm. Interviewers have good radar for candidates who are going through motions versus candidates who actually want to be there. If you're not sure whether you want this job, no prep workflow will paper over that ambivalence. But if you do want it, AI can help you show that clearly — by removing the friction of underprepared answers and leaving you free to be present in the conversation.
And finally: AI works from what you give it. If you paste in a job description and your background and ask for specific help, you'll get specific help. If you just ask "what questions should I expect?", you'll get generic advice that could apply to anyone. The more context you provide, the more useful the output.
Think of it as a very patient, very available research and practice partner — one that doesn't mind running through the same question five times until your answer is clean, and doesn't get tired of your nervousness about the interview. That's the actual value. Use it that way.
— Tom