I want to talk about the uncomfortable thing structured interviews do, which is hold a mirror up to the interviewer. Most advice about hiring bias is aimed at the obvious stuff: don't ask about kids, don't penalize an accent, don't assume the quiet candidate is less capable. All real, all useful, but it's not where most of my bias actually lived. Mine lived in the halo effect — a candidate nails the first question, and now everything else they say sounds smarter than it is, and I spend the rest of the interview gathering evidence for a verdict I reached in the first three minutes.
The experiment I ran on myself
I generated an interview guide with the Structured Interview Questions & Scoring Rubric Generator prompt using Claude Sonnet 4.6 for a customer success role: four competencies, one final-panel stage. Then I interviewed two candidates back to back with the identical questions and the identical rubric, and I forced myself to write one quote of evidence per score before I wrote the score itself. That last constraint — evidence first, number second — is the entire trick, and it's the part of the prompt's scorecard I used to skip because it felt like busywork.
Candidate A charmed me in the first three minutes. Warm, quick, told a great story about a save. Candidate B was nervous, spoke slowly, and took a long beat before each answer. My gut had ranked them within ninety seconds, and my gut was very confident. I've been interviewing for years; I trust my gut; that confidence is precisely the problem the rubric is designed to interrupt.
It's worth saying that nervousness is not a competency. A slow, careful speaker can be a phenomenal customer success manager, and a fast, charming talker can be a disaster who over-promises and torches renewals. But interviews reward performance-under-interview-conditions, which is a skill almost entirely unrelated to the actual job. The rubric's whole job is to keep dragging my attention back to what the person actually did at work, not how smoothly they're doing the artificial task of being interviewed by me.
What the evidence actually said
When I went competency by competency, writing the evidence quote first, Candidate B had given measurably stronger answers on two of the four competencies — a concrete churn number she'd reversed, a specific escalation she'd defused with a named tactic, a renewal process she'd built from scratch. Candidate A had charisma and, on those same two competencies, surprisingly little underneath: warm generalities, a story that fell apart under the follow-up probe. The rubric did not care how either of them made me feel in the first three minutes. It only cared what they actually said when asked the same questions.
- Write the evidence quote before the score — it stops you backfilling a number you already decided
- Score each competency in isolation, not as one overall gut impression
- If two candidates 'feel' very different but score close, your feeling is doing more work than the evidence
- Keep the weak-answer red flags from the prompt visible while you interview — they catch the rehearsed story before it charms you
Why behavioral questions matter more than I thought
The prompt insists on STAR behavioral questions and bans hypotheticals, and I used to think that rule was pedantic. It isn't. "What would you do if a customer churned?" gets you a tidy, idealized answer the candidate assembled in the shower that morning — it tests imagination, not experience. "Tell me about a time you lost an account and what you did about it" gets you the truth, including the parts where it went badly and what they actually learned. The follow-up probe the prompt generates for each question is where the real signal hides, because it's the one part of the conversation nobody rehearses. Candidate A's polished story had no second act; the probe found the edge of what he'd really done.
The takeaway
Structured interviewing didn't make me a colder interviewer. It made me an honest one. I still notice charisma — I'm human, and charisma is a real and valuable skill in a customer-facing role — I just don't let it quietly inflate four unrelated competency scores anymore. The feeling still happens; the rubric just stops it from secretly doing all the voting.
If you've ever walked out of an interview certain about someone and then struggled to articulate why, that uncertainty is your bias quietly asking for a rubric. The fix isn't to become suspicious of your instincts — it's to make your instincts show their work. Grab the Structured Interview Questions & Scoring Rubric Generator prompt on Prompt Dock, interview your next two candidates with the exact same guide, and write the evidence before the number every single time. You'll learn more about your own judgment than about either candidate — and you'll hire the one who can actually do the job rather than the one who's best at the first ninety seconds of meeting you. I run it alongside my Inclusive Job Description Writer from a One-Line Role Brief prompt so that bias gets caught at both ends of the funnel, on the way in and on the way through.