Structured interview guides and scoring rubrics that reduce bias
Unstructured interviews are inconsistent and legally exposed: different candidates get different questions and gut-feel scores. HR can use AI to build a structured kit — the same job-related questions for every candidate plus an anchored 1-5 rubric — which improves hiring quality and creates the consistent, defensible record that EEOC guidance favors. It also keeps interviewers away from questions that touch protected characteristics.
You are an HR assistant building a structured interview kit. Base everything on the job requirements I provide — do not assume duties I did not list. Role: {{job_title}} Key competencies to assess: {{competencies}} Interview length and panel: {{format}} Produce: 1. 6-8 behavioral and situational questions, each tied to one competency and phrased to draw out a specific past example (STAR-friendly). Job-related only. 2. For each question, 2-3 follow-up probes. 3. A scoring rubric: for each competency, a 1-5 scale with a concrete behavioral anchor describing what a 1, a 3, and a 5 answer look like. 4. An interviewer note listing question types to avoid (anything touching age, family or marital status, health or disability, religion, national origin, or arrest record) and why they are off-limits. Rules: - Every question must be job-related and legally safe. Do not generate questions about protected characteristics or that act as a proxy for them. - Do not predict or assume how a candidate would answer. - Keep it practical for a {{format}} interview. - This kit supports human interviewers; final scoring and the hiring decision are made by people.
Fill in your details and the prompt updates live — then copy.
Competency: Handling escalations Q3 (situational): "Tell me about a time a customer was upset about something that wasn't your fault. What did you do?" Probes: What options did you weigh? How did the customer respond? What would you do differently? Rubric — Handling escalations: 1 — Blames the customer or others; no ownership. 3 — De-escalates and resolves the immediate issue. 5 — Resolves, follows up, and proposes a fix so it doesn't recur. Avoid: "Do you have kids at home?" — family-status questions aren't job-related and create discrimination risk.
The full workflow
- Pull the real competencies straight from the job description
- Run the prompt and sanity-check every question for job-relevance and legal safety
- Have all interviewers use the same questions and score against the anchors
- Keep the completed rubrics as your hiring record; make the decision as a human panel
Watch out for
AI can suggest questions that are unlawful or a proxy for protected traits. You are responsible for screening them out — review against current EEOC guidance before use.
AI scores; it does not decide. Automated rejection or ranking tools face disparate-impact liability and bias-audit laws (for example NYC Local Law 144 and Illinois rules), and the employer is liable even when a vendor built the tool. A human makes the hire or no-hire call.
Never paste real candidates' resumes, names, or interview notes into a consumer AI account — keep candidate data in your ATS or an approved enterprise tool.
Where this comes from
Every use case on this site is grounded in real reports from working hr managers — not invented by us.