When you ask a candidate to write interview questions with ChatGPT, everyone teaches them to prepare. This article teaches you — the person conducting the interview — to build a system that makes gut-feeling hiring obsolete.

If you’re the hiring manager for a role, you’ve almost certainly sat down the morning of an interview and cobbled together your questions from whatever came to mind, the candidate’s resume, and your vague memory of what you asked the last person. That’s an unstructured interview, and the evidence on unstructured interviews is fairly consistent: they’re worse than a coin flip for predicting performance, largely because they measure how well someone interviews rather than how well they’ll do the job.

Here’s what actually matters about learning how to write interview questions using ChatGPT: the goal isn’t to generate questions — it’s to build a complete evaluation system. That means competency extraction, behavioral question generation, an objective grading rubric defining what a strong answer actually sounds like, and a way to coordinate a panel so four interviewers cover different ground. This article gives you the exact prompt sequence to do all of that in under five minutes.

This article sits at a different point in the hiring funnel from our earlier cluster pieces — if you’re still writing the job description, see our guide on writing a job description using AI first. And once you’ve hired someone, our guide on writing a new hire welcome message picks up from here.

Why “Tell Me About Yourself” Is Costing You Great Hires

A common mistake is treating interview prep as a one-time task rather than a repeatable system. The result is that most interviews are actually pretty inconsistent: Candidate A gets asked about a specific project challenge, candidate B gets asked a generic “what’s your biggest weakness,” and the two interview experiences are so different that comparing the results fairly is nearly impossible.

The reality is unstructured interviews — where interviewers ask whatever comes to mind — have a well-documented problem with hiring bias. Questions like “what’s your biggest weakness” and “where do you see yourself in five years” are almost always answered with polished, rehearsed responses that reveal nothing about actual on-the-job behavior. What you actually want to know is what a person *did* in a situation analogous to the challenges they’ll face in this role, not what they’re planning to say when asked about their ambitions.

What many people overlook is that the consistency problem is also a legal problem. Asking different candidates different questions means you’re not evaluating them on the same criteria, which creates exposure if a hiring decision is ever challenged. SHRM’s guidance on the importance of structured interviews in reducing bias consistently points to standardized question sets as the single most effective tool for fairer, more defensible hiring.

The 4-Step Workflow to Write Interview Questions Using ChatGPT

In practice, writing interview questions is a four-step process — not a single prompt. Each step builds on the previous one, and the system is what makes the output genuinely useful rather than just a list of reasonable-sounding questions.

Step 1: Extract Core Competencies From the Job Description

Before you can generate meaningful questions, you need a clear picture of what you’re actually testing. Paste the job description and ask the AI to do the analytical work of identifying what skills and behaviors matter most — not just what the JD says, but what those skills look like in practice.

Act as an expert Corporate Recruiter. I am pasting a job description for a [Job Title] role. Analyze the text and extract the top 5 core competencies required for this role.

For each competency, write:
- A one-sentence definition of what it looks like in practice on this specific job
- One example of a daily situation where this competency would be tested

Job Description:
[PASTE JOB DESCRIPTION HERE]

Step 2: Generate Behavioral Questions (The STAR Method)

A candidate can memorize an answer to “describe your leadership style.” They cannot fake a detailed, specific answer to “tell me about a time you led a team through a project that was about to miss a major deadline.” The STAR method (Situation, Task, Action, Result) forces the candidate to describe a real past event — which means their answer either has the texture of lived experience or it doesn’t.

Based on the competency "[INSERT COMPETENCY — e.g. Stakeholder Management]" required for this [Job Title] role, generate 3 behavioral interview questions that require the candidate to use the STAR method.

CRITICAL: Avoid generic questions. Make the scenarios highly specific to the daily realities of a [Job Title] working in the [Industry] sector.

Also provide: for each question, the ideal follow-up probe I should ask if the candidate gives a vague or overly polished answer.

Step 3: Craft Scenario-Based Technical Questions

This becomes important when you’re hiring for a role you don’t fully understand yourself. Behavioral questions test how someone handled a past situation; scenario-based questions test how they’d reason through a specific, current challenge. For technical roles especially, scenario questions are harder to bluff through because they require the candidate to think out loud in front of you.

I am a non-technical manager interviewing candidates for a [Technical Job Title] role. I need to assess their actual competence, not just their buzzwords.

Give me 3 scenario-based questions to ask. For each question:
1. Tell me exactly what a correct, competent answer sounds like in plain English
2. List 2 specific "red flags" or jargon-heavy non-answers that indicate the candidate is bluffing
3. Provide the follow-up question I should ask to test whether they actually know what they're talking about

Step 4: Building the “Listen-For” Candidate Evaluation Rubric

This is the step most interviewers skip entirely — and the most valuable one. You ask a great behavioral question, the candidate gives a confident-sounding answer, and then you have no idea whether what you just heard was an excellent response or an average one dressed up in good delivery. A listen-for rubric solves that by defining in advance what a poor, average, and excellent answer actually includes.

For the following interview question: "[PASTE YOUR QUESTION]"

Create a grading rubric for the interviewer. Define exactly what a 1-star (Poor), 3-star (Average), and 5-star (Excellent) answer sounds like in plain language.

For the 5-star answer, provide 3 specific "Listen For" bullet points — the exact evidence, keywords, or elements that would indicate a top-tier candidate rather than a well-rehearsed average one.

3 More ChatGPT Prompts for Interview Questions You Can Copy Today

The four-step workflow above covers the core system. These three additional prompts fill the specific gaps that trip up most hiring managers: culture fit questions that don’t invite discrimination lawsuits, a way to detect candidate bluffing in real time, and a compare-card showing the difference between a weak and strong prompting approach.

The Culture and Core Values Prompt

Culture fit questions are where interviewers accidentally introduce discrimination — “I’m just looking for someone who fits our culture” is not a legally defensible reason for a hiring decision. The prompt below forces the AI to frame culture questions around observable, role-relevant behaviors rather than personality traits or subjective “fit.”

I want to assess whether a candidate aligns with our company's core values. Our top 3 values are: [VALUE 1], [VALUE 2], [VALUE 3].

Write 1 behavioral interview question for each value. Each question must require the candidate to describe a real, past work situation — not share an opinion.

GUARDRAIL: Do not write questions about religion, family status, national origin, political beliefs, or any other protected characteristic. Focus entirely on observable, professional behaviors.

The Red Flag Detection Prompt

When you ask a behavioral question and get a smooth, polished answer, you can’t always tell if you’re hearing genuine experience or a well-rehearsed story. The follow-up question is what separates them — but most interviewers don’t know which follow-up will actually reveal the gap.

I just asked a candidate: "[YOUR BEHAVIORAL QUESTION]" and they gave me this answer: "[BRIEFLY SUMMARIZE THEIR ANSWER]"

Analyze whether this answer demonstrates genuine experience or could be fabricated. Then:
1. Provide 1 specific follow-up question to probe deeper and verify the story is real
2. Tell me the exact phrase or detail I should listen for that indicates authentic experience vs. vague rehearsal

How to Coordinate a Panel Interview Using AI

One of the most consistently complained-about interview experiences from candidates — and one of the most overlooked by hiring teams — is arriving at a 4-person panel only to be asked “walk us through your background” by all four of them. It wastes everyone’s time, frustrates the candidate, and produces four almost-identical data points instead of the 360-degree picture that’s actually the point of a panel.

I am organizing a 4-person panel interview for a [Job Title] role. My 4 interviewers are: the Hiring Manager, a Peer from the same team, a Direct Report (if applicable), and the HR Director.

Divide the interview strategy across these 4 people so there is zero overlap in topics covered. For each interviewer, provide:
- Their specific focus area (1 sentence)
- 3 tailored questions aligned to that focus
- 1 follow-up probe per question to use if the candidate's initial answer is vague

Job description context: [PASTE KEY ROLE DETAILS]

For Microsoft 365 teams who want to move the finished interview guide directly into a clean, formatted document, formatting your candidate evaluation sheets in Word with Copilot covers the native workflow. Teams working in Google Workspace can easily export your interview rubrics into Google Docs using the same process.

Will Candidates Know You Used AI? (And Does It Matter?)

The hidden anxiety for most hiring managers is that using AI to generate interview questions feels like it undermines the authenticity of the process. In practice, this is backwards. The AI generates the structure; you supply the context — the specific role, the specific challenges this person will face, the specific competencies that actually matter on your team. A well-used AI prompt doesn’t produce generic questions; it produces better-structured, more specific questions than most managers would write manually in the time available.

What candidates actually notice isn’t whether questions were AI-generated. What they notice is whether the interviewer read their resume before the call started, whether the questions were thoughtful and role-specific, and whether the whole experience felt like someone actually cared about finding the right person rather than going through the motions. All of those things improve when you use the system in this guide.

One genuine limitation worth naming: AI will sometimes produce questions that are technically well-structured but slightly off-tone for your company’s culture, or that address a skill at a level of seniority that doesn’t quite match the role. Always read the output before using it. The productivity consultant’s principle applies: AI provides the framework; you must provide the context and the final judgment.

Frequently Asked Questions

Next Steps