How to Write Interview Questions That Actually Reveal Something

Most interviews tell you who is good at interviews. “What’s your greatest weakness?” and “Where do you see yourself in five years?” have rehearsed answers and reveal almost nothing. When managers ask a chatbot for interview questions, they usually get exactly that list of tired classics. A good hiring manager builds interviews that surface evidence, not performance — and the questions are only half of it.

The skill behind this is Blake, a hiring persona you load once into Claude, ChatGPT, or any AI chat. Here is how it builds an interview that actually discriminates between candidates.

Start from the scorecard, not the questions

The mistake is writing questions first. Blake starts from what the role actually requires — the few competencies that predict success — and designs questions to test each one. If a question is not measuring something on the scorecard, it does not earn a slot.

Behavioural over hypothetical

“What would you do if…” invites a candidate to describe their best self. “Tell me about a time you actually did…” asks for evidence. Blake writes behavioural questions that dig into real past situations, with the follow-up probes that separate a real story from a polished one.

A way to score that is not a gut feel

Blake builds a simple rubric for each competency — what a weak, solid, and strong answer sounds like — so two interviewers can compare notes on the same scale instead of trading vibes. Structured scoring is the single biggest upgrade most hiring processes can make.

Questions that reduce, not amplify, bias

Asking every candidate the same role-relevant questions, scored the same way, is fairer than a free-flowing chat that rewards similarity to the interviewer. Blake keeps the interview on job-relevant ground and steers away from the small talk where bias quietly creeps in.

How to get the most out of it

Give it the role and the two or three things that genuinely predict success in it. Ask for the scorecard first, then questions mapped to it, then the probes. And keep the decision human — the skill builds a fairer, sharper process; the judgement is still yours.

It works with Claude, ChatGPT, or any AI chat that accepts a system prompt. For the rest of hiring, the work & business skills collection covers screening, recruiting, and onboarding — each a focused assistant rather than a general chatbot.