ChatGPT interview prep prompts for a realistic mock interview

Most people prepare for a job interview by reading a list of common questions and nodding along in their head. That feels like practice, but it isn't. The hard part of an interview is answering out loud, in order, without knowing what's coming next.

An AI chat tool can fill that gap. It can play the interviewer, ask one question at a time, wait for your answer, and then tell you what was clear and what wandered. It's available at 11 p.m. the night before, it never gets bored, and nobody else has to hear your first clumsy attempt.

The trick is setting it up properly. If you just ask for "interview questions", you get a list of 20 questions you'll skim and forget. Below is the difference.

Bad vs. good: the mock interview prompt

Before you start: gather your material

The AI can only tailor the practice if you give it something to work with. Have these ready to paste:

• The job posting, especially the responsibilities and "what we're looking for" section.
• Your résumé, or a short list of your relevant jobs and accomplishments.
• Anything you know about the interview: who's interviewing you, how long it is, whether it's a phone screen or a panel.

Before pasting your résumé, it's sensible to remove your home address, phone number, and email. The AI doesn't need them to help you practice. If you haven't tuned your résumé for this job yet, our guide to AI prompts for your résumé walks through that first.

The main mock interview prompt

This is the prompt to save. It sets the role, the pace, and the kind of feedback you want.

Two lines in that prompt do most of the work. "Stop and wait for my answer" keeps the AI from asking and answering everything at once, which it tends to do otherwise. "Don't just praise me" matters because AI tools lean polite; without permission to be critical, feedback can be uselessly nice.

Why this works

A good prompt gives the AI a job, a situation, and rules. Here, the job is "interviewer", the situation is your actual posting and background, and the rules control the pace and the feedback. Giving the AI a role like this is called role prompting, and our role prompting guide explains why it changes the answers you get. The other key is going one question at a time: you practice thinking on your feet instead of reading ahead.

Build your "tell me about a time" stories

Behavioral questions ("Tell me about a time you handled a conflict") trip up a lot of people because they need a real story, told in a clear order. A common structure is STAR: the Situation, the Task you had, the Action you took, and the Result. The AI is good at helping you shape a rough memory into that structure, as long as the facts come from you.

That last instruction is the important one. If the AI invents a result that sounds impressive, you might repeat it in the interview and then be asked a follow-up you can't answer. Our guide on how to stop ChatGPT making things up has more ways to keep it honest.

Practice the questions you're dreading

Everyone has one question they hope doesn't come up: why you left your last job, a gap in your work history, a job you didn't stay at long, or a missing qualification. Practicing that one answer often helps more than practicing ten easy ones.

Prepare questions to ask them

Near the end of most interviews, you'll be asked, "Do you have any questions for us?" Having two or three good ones shows you've thought about the role.

Quick-fire practice when you're short on time

If the interview is in an hour and you just want a warm-up, use a short version.

Practice out loud, not just in writing

Typing answers is useful, but interviews are spoken. Two ways to make practice more realistic:

• Use voice mode. ChatGPT and Claude both offer voice conversations in their mobile apps at the time of writing, so you can answer out loud. Features change, so check your app.
• Type, then say it. Type your answer for feedback, then say the improved version out loud two or three times. It's the speaking that makes it stick.

ChatGPT vs Claude for interview practice

Both work well for mock interviews, and the prompts above work the same way in either. In our experience, the differences are small and depend more on the prompt than the tool. Either one may run ahead and ask several questions at once, or soften its feedback; the fixes in this guide handle both. ChatGPT's voice mode is widely used for spoken practice, and Claude also offers voice in its apps; try whichever you already have an account with. Neither knows what a specific company really asks in interviews, so treat any "this company typically asks" claim with caution and check sources like the company's own careers page. For a broader comparison, see ChatGPT vs Claude for beginners.

Useful follow-up prompts

• "That feedback was too gentle. Be more direct about what a hiring manager would think."
• "Ask me a follow-up question to my last answer, the way a real interviewer would dig in."
• "My answer was too long. Show me a version that's half the length."
• "Switch roles: you're now a tougher, more skeptical interviewer. Ask the next question."
• "Summarize the three things I should practice most before the real interview."

More ideas for steering a conversation are in our guide to AI follow-up prompts.

Want more ready-to-use prompts for work, home, and everyday tasks? Get 50 prompts like this free (PDF).

A few honest cautions

Don't memorize a script. Use the AI to find your main points, not to write a speech. Answers that sound recited can come across as stiff.

The AI isn't the hiring manager. Its feedback is a useful outside view, not a prediction. Real interviewers vary a lot.

Keep it true. Everything you say in an interview should be something you can back up. If the AI polishes a story past what happened, pull it back.

Once the interview is done, a short thank-you note helps, and our work email prompts can help you write one in a couple of minutes. If you're still at the application stage, start with our cover letter prompts.

Frequently asked questions