Day 2: Generate Structured Interview Questions
The Concept
Ask most hiring managers how they run interviews and they will describe a conversation. Ask them what they are specifically trying to assess in that conversation and many will pause. The relationship between what is discussed and what is actually measured is, in most unstructured interviews, loose at best. This is not a character flaw — it is a structural problem. Without a defined set of competencies, pre-set questions, and a common scoring framework, every interviewer is running a different process. And different processes produce inconsistent, hard-to-defend hiring decisions.
The research on this point is not ambiguous. Structured interviews — those that use the same questions, in the same order, scored against defined criteria — consistently outperform unstructured interviews in predicting job performance. The predictive validity of structured interviews is roughly twice that of unstructured ones, according to decades of meta-analytic research. Yet most organisations continue to leave interview design to individual hiring managers, with results that vary as much as the managers themselves.
The Evidence for Structure
The reason structured interviews work is straightforward: they hold the process constant so you are comparing candidates against a common standard rather than against each other's conversational charm. When two interviewers ask the same question and score responses against the same rubric, their combined assessment is more reliable than either one conducting a free-flowing chat. When each interviewer asks different questions, you end up comparing apples and oranges in a debrief — and the most persuasive voice in the room tends to win, which is a function of seniority, not insight.
Behavioural questions — those that ask candidates to describe specific past situations — are the workhorse of structured interviewing because past behaviour is a better predictor of future behaviour than hypothetical intent. "Tell me about a time when..." forces candidates to draw on real experience rather than describe their best self. Situational questions add a complementary layer by testing judgement against a scenario, which is particularly useful for roles where candidates may lack direct experience but bring transferable skills.
Why Most Organisations Still Wing It
If the evidence for structured interviews is so clear, why do so few organisations use them systematically? The answer is mostly friction. Designing a proper question bank per role takes time that most hiring managers and recruiters do not have. Calibration — the process of aligning all interviewers on what a good answer looks like — requires a meeting that rarely makes it onto the calendar. And once a hire is done, the question bank sits in a document that nobody revisits for the next round.
AI removes the friction from the design phase entirely. Generating a full question bank with behavioural, situational, and probe questions for five competencies used to take several hours of careful work. With a well-structured prompt, it takes minutes — producing a first draft that a skilled HR professional can review and refine rather than build from scratch. The calibration conversation still needs to happen. The question bank now gives you something concrete to calibrate around.
Using the Bank to Start the Real Conversation
The most important thing you do with the AI-generated question bank is not use it as-is. It is use it as the foundation for a conversation with the hiring manager. Print it out, sit down together, and ask: which of these questions would actually reveal what we need to know about this candidate? Which ones feel too generic for this specific team? Are there scenarios particular to our current challenges that would be more revealing than the standard ones?
That conversation — structured around an existing draft rather than a blank page — is where real hiring quality is built. The AI handles the scaffolding; you and the hiring manager construct the process on top of it. By the end of today's exercise, you should have a question bank you are confident enough to hand to any interviewer on the panel, knowing that whoever runs the interview, the process is consistent enough to make the debrief a fair comparison.
Prompt of the day
Copy this into your AI tool and replace any bracketed placeholders.
Prompt
You are an organisational psychologist who designs structured interview frameworks for hiring teams. I am hiring for the role of [JOB TITLE] at [COMPANY NAME]. The key competencies required for success in this role are: [LIST 3-5 COMPETENCIES, e.g. 'stakeholder management, analytical thinking, influencing without authority']. Please generate a structured interview question bank with the following: 1. For each competency, write two behavioural questions (using the STAR format — Situation, Task, Action, Result), one situational question (a hypothetical scenario), and one follow-up probe question that goes deeper on a vague answer 2. Write a two-sentence role overview I can read aloud at the start of the interview to set context 3. Write three questions the candidate is likely to ask, with a one-paragraph suggested response to each 4. Flag any competencies where bias commonly enters interview scoring, and suggest how to mitigate it Format the output as a table for the question bank. Plain text for the overview and candidate questions.
Your 15-minute task
Choose a role you are actively recruiting for or have hired for recently. Identify three to five competencies that genuinely predict success in the role — not generic ones, but specific to this team and context. Fill in the placeholders and run the prompt. Then share the question bank with the hiring manager and ask them to mark which questions they would actually use.
Expected win
A ready-to-use structured interview question bank with behavioural, situational, and probe questions for each competency, a candidate-ready role overview script, and a bias flag section — all formatted so the hiring manager can pick it up and run an interview today.
Power user tip
After the hiring manager reviews the question bank, send this follow-up to your AI tool: 'Now create a simple scoring rubric for each competency. For each one, describe what a 1-out-of-5 answer looks like, what a 3-out-of-5 answer looks like, and what a 5-out-of-5 answer looks like — based on the questions you just wrote.' A rubric transforms a good question bank into a calibrated panel that scores consistently across interviewers.