Quick answer: A fair AI interview starts with a clear prompt. In Mitoo Hire, include seven things in every AI interview prompt: the job role, required skills, experience level, interview questions, question difficulty, scoring criteria and follow-up instructions. When these are set once per role, every candidate gets job-related questions at the same level and is scored against the same standard.
Take an HR manager at a growing business who posts one opening for an Inside Sales Executive and receives 250 applications in a week. She shortlists 40 for a first-round call. By the tenth call, her questions have drifted. Some candidates are asked about CRM tools, others are not. One gets an easy warm-up, another gets a tough objection-handling scenario. At the end, comparing them fairly is almost impossible.
AI first-round interviews fix the time problem. But the questions the AI asks are only as good as the prompt you give it. A vague prompt like “Interview this candidate for sales” leads to generic questions and uneven scoring. A well-written prompt gives every candidate the same fair chance.
This guide walks you through the seven things to include in your AI interview prompt, with a ready-to-use template you can adapt in Mitoo Hire.
Why your AI interview prompt decides how fair the interview is
An AI interviewer follows your instructions closely. If the prompt is clear, the questions are clear. If the prompt leaves gaps, the AI fills them in its own way, and that can change from one candidate to the next.
A strong prompt turns a loose chat into structured AI interview questions and helps your team in three ways:
- Fairness: Every candidate for the same role is asked questions on the same topics, at the same level.
- Consistency: Answers are scored against the same criteria, so a score of 7 means the same thing for every candidate.
- Faster shortlisting: Recruiters compare candidates side by side instead of re-reading notes from 40 different conversations.
Think of the prompt as your interview brief. You would not send a new panel member into an interview without a job description and a scoring sheet. The AI needs the same.
7 things to include in your AI interview prompt
1. Job role
Start with the exact job title and a two-to-three line summary of what the person will do day to day. “Sales Executive” means different things in a B2B software company and a retail chain. Add the department, location or work mode, and who the role reports to.
Example: “Inside Sales Executive, B2B team, Pune office. Handles inbound enquiries, qualifies leads and books demos for the field sales team.”
2. Required skills
List the skills the role truly needs, split into must-have and good-to-have. Keep it to five or six must-haves so the AI stays focused. This stops the interview from drifting into topics that do not matter for the job.
Example: “Must-have: clear spoken English and Hindi, lead qualification, CRM data entry, handling objections. Good-to-have: experience selling to manufacturers.”
3. Experience level
Tell the AI who it is speaking to: a fresher, someone with two to four years of experience, or a team lead. The same question about “handling a difficult client” should sound very different for a fresher and a manager. Setting the level keeps expectations fair and realistic.
4. Interview questions
Give the AI a core set of questions that every candidate must answer, usually five to eight for a first round. Mix question types:
- Role-based: “Walk me through how you qualify a new inbound lead.”
- Behavioural: “Tell me about a time you missed a sales target. What did you do next?”
- Situational: “A customer says your price is 20% higher than a competitor. How do you respond?”
A fixed core set is the single biggest step towards consistency. Candidates can still get different follow-ups, but everyone answers the same core questions.
5. Question difficulty
Say how hard the questions should be and how the interview should flow. A good pattern is to start easy, move to role-specific questions, and end with one scenario that tests judgement. Match difficulty to the experience level from point 3, and ask the AI to keep it the same for every candidate.
6. Scoring criteria
This is where most prompts fall short. Tell the AI exactly what to score and what a good answer looks like. Use a simple 1 to 5 scale with a short description for each level.
| Criterion | Weight | What a 5 looks like | What a 1 looks like |
| Role knowledge | 30% | Explains lead qualification steps clearly with examples | Vague or unrelated answer |
| Communication | 25% | Clear, structured, easy to follow | Hard to follow, no structure |
| Problem-solving | 25% | Handles the price objection with a reasoned approach | Gives up or offers a discount straight away |
| Relevant experience | 20% | Gives specific past results | No examples given |
Clear criteria mean two recruiters reviewing the same report will reach the same conclusion.
7. Follow-up instructions
Follow-ups help the AI dig deeper when an answer is too short or vague. Set clear rules so follow-ups stay fair:
- Ask at most one or two follow-ups per question.
- Only ask follow-ups linked to the skills in point 2.
- Do not ask about age, marital status, religion, caste, family plans or other personal details.
- If a candidate does not know an answer, move on politely without penalising tone.
These rules keep the conversation natural for the candidate and keep the interview focused on the job.
A ready-to-use AI interview prompt template
Copy this template into Mitoo Hire and fill in the brackets for each role.
Role: [Job title], [department], [location / work mode]. Reports to [manager role].
Role summary: [2–3 lines on daily work]
Must-have skills: [skill 1], [skill 2], [skill 3], [skill 4], [skill 5]
Good-to-have skills: [skill 1], [skill 2]
Experience level: [Fresher / 1–3 years / 3–6 years / Team lead]
Core questions (ask every candidate, in this order):
1. [Warm-up question about background]
2. [Role-based question]
3. [Behavioural question]
4. [Situational question]
Difficulty: Start easy, then role-specific, end with one scenario. Keep difficulty the same for every candidate.
Scoring (1–5 per criterion):
- Role knowledge (30%): [what a 5 looks like]
- Communication (25%): [what a 5 looks like]
- Problem-solving (25%): [what a 5 looks like]
- Relevant experience (20%): [what a 5 looks like]
Follow-up rules:
- Max 2 follow-ups per question, only on listed skills.
- No personal questions (age, marital status, religion, caste, family plans).
- If the candidate is unsure, move on politely.
- End by asking if the candidate has any questions about the role.
Common mistakes to avoid
- Writing a one-line prompt. “Interview for a sales role” gives the AI too much freedom, and questions change from candidate to candidate.
- Listing too many skills. Fifteen must-haves spread the interview thin. Pick the five or six that matter most.
- Skipping scoring criteria. Without them, scores are hard to compare and harder to explain to hiring managers.
- Mixing levels. Asking a fresher manager-level questions is unfair to the candidate and wastes a good profile.
- Leaving follow-ups open-ended. Unlimited follow-ups can turn one candidate’s interview into a grilling while another’s stays light.
- Never reviewing the prompt. Check the first few interview reports for each role and tighten the prompt if questions or scores look off.
How Mitoo Hire makes this easier
Mitoo Hire, an AI recruitment tool, brings your first-round screening into one place. Instead of rewriting instructions for every candidate, you set up the AI interview once per role and reuse it for every applicant.
- One prompt per role: Save the job role, skills, experience level and questions together, so every candidate for that opening gets the same interview.
- Consistent scoring: Candidates are evaluated against the same criteria, making shortlists easier to compare and explain.
- Controlled follow-ups: Your follow-up rules keep conversations natural without letting them drift off the job.
- Organised screening: Recruiters review results in one place instead of chasing notes and call recordings.
The result is consistent candidate screening that is faster for your team and fairer for every candidate.
Frequently asked questions
What is an AI interview prompt? It is the set of instructions you give the AI interviewer. It tells the AI which role it is hiring for, what to ask, how hard the questions should be and how to score answers.
How many questions should a first-round AI interview have? Five to eight core questions work well for most roles. That is enough to judge fit without making the interview too long for candidates.
Can every candidate get exactly the same questions? Yes. Add a fixed set of core questions and ask the AI to use them in the same order. Follow-ups may differ slightly based on answers, but your follow-up rules keep them within limits.
How do scoring criteria make hiring fairer? They give every answer the same yardstick. Recruiters can see why one candidate scored higher than another, which makes shortlisting decisions easier to explain.
Should I use one prompt for all roles? No. Create a separate prompt for each role, or at least each role family. A prompt for an accountant should not look like one for a field sales executive.
Conclusion
Fair hiring does not start in the interview room. It starts with the instructions you write before the first candidate logs in. When your AI interview prompt covers the job role, required skills, experience level, interview questions, question difficulty, scoring criteria and follow-up instructions, every applicant gets the same fair chance, and your team gets results it can trust.
Start with your highest-volume role, set up the prompt using the template above, and test it with a small batch of candidates. Then refine and roll it out across your other openings.
Ready to run fairer first-round interviews? Explore Mitoo Hire and book your free demo today.
Call: +91 70418 00957
Email: info@mitoo.co.in
Website: https://mitoo.ai
Disclaimer: The information provided is for general information purposes only. It may vary over time, so we recommend verifying the details independently before making any decisions.