Create consistent job-related questions that elicit evidence without revealing a preferred answer
9 min setupTested with:ChatGPTReviewed: 2026-08-28
1
Add your context
Your text stays in this browser. AILesson Prompts does not send it to a model or server.
2
Your prompt
Unfilled fields remain visible as placeholders, so you can still copy and edit the prompt
Generate a structured, non-leading interview guide from approved job criteria.
Duties, essential criteria, evidence anchors, trainable skills, conditions, and exclusions:
[criteria]
Stage, duration, interviewers, instructions, response modes, work sample, scoring, and accommodations:
[format]
Jurisdiction, prohibited topics, privacy, consistency, follow-up limits, notes, bias controls, and approvers:
[governance]
Map every question to exactly one scored criterion and explain the job relevance. Use open behavioral, situational, or work-sample questions that do not contain the desired method, praise a presumed trait, assume past access or privilege, or ask candidates to agree with a premise. Provide the same primary questions and equivalent instructions to all candidates. For each include neutral clarification and depth probes, observable evidence indicators, insufficient-evidence handling, and topics that must not influence scoring. Do not ask directly or indirectly about protected or unrelated personal information, salary history, family, health, age, religion, nationality beyond approved work-right handling, or speculative future personal plans. Offer equivalent response routes for approved accommodations without lowering criteria. Fit questions, candidate answers, and transitions into the duration. End with order, timing, interviewer script, consistency log, legal/HR review flags, and a test that removes clues to the preferred answer.
Private by defaultPrompt assembly happens locally in your browser. Avoid placing confidential information into any AI service unless your organization allows it.
From input to outcome
A worked example
See how concrete context turns this recipe into a usable result
Actual input
Approved selection criteria
Operations Coordinator criteria: organizing concurrent requests; clear written communication; improving a recurring process; protecting routine operational data and preserving decision authority. Linear is trainable and not scored. Anchors require concrete situation, actions, evidence, and trade-offs. No degree or years requirement.
Interview format
45-minute first interview, four questions, Hiring Manager leads and Support Lead notes. Rating 1-4 plus Not enough evidence. Candidate may answer orally or in writing and use arranged assistive technology. No live work sample in this stage. Five minutes for candidate questions.
Governance rules
UK; HR prohibited-topic guide controls. Do not ask about family, health, age, salary history, nationality beyond separate work-right process, or personal schedule beyond ability to meet approved 13:00-17:00 UTC overlap. Ask all candidates same primary questions; neutral probes may differ only to clarify evidence. Record evidence, not personality. HR approves before use.
Example output
Opening script, 3 minutes: explain the role-relevant criteria, four-question structure, note-taking, response options, and that candidates may ask for a question to be repeated. Do not mention a preferred method.
Concurrent requests, 8 minutes: “Tell us about a specific time you had several requests or pieces of work moving at once. How did you decide what information to track and what to do next?” Neutral probes: “What was your responsibility?” “What changed during the work?” “What evidence showed the system was or was not working?” Do not ask “How did you use a priority matrix?” Evidence: volume, dependencies, ownership boundaries, adaptation, outcome.
Written communication, 7 minutes: “Describe a written update you created for people with different information needs. What did each reader need, and how did you check the message worked?” Probes seek artifact, audience decisions, factual checks, and feedback. Do not score vocabulary sophistication, accent, or access to prestigious settings.
Process improvement, 9 minutes: “Walk us through one recurring process you changed. What was happening before, what did you change, and how did you evaluate it?” Neutral probes: baseline evidence, people involved, trade-off, later revision. The question does not assume the change succeeded.
Data and authority judgment, 8 minutes: “Tell us about a time information needed to move between people but some details or decisions were not yours to share or make. How did you handle it?” Probe only for applicable boundaries, escalation, and evidence. Do not solicit confidential details.
Reserve 5 minutes for candidate questions and 5 minutes for transitions/overrun. Mark Not enough evidence rather than infer a score. Consistency log records primary question, only probes used, response mode, accommodation implementation, and evidence locator. HR must review prohibited-topic wording and the question about schedule separately if it is added later.
Why this works
1
Single-criterion questions make candidate evidence easier to compare without hidden weighting.
2
A preferred-answer clue test exposes questions that coach one candidate response.
Check the result
Does each question assess one approved job criterion without embedding the answer?
Are primary questions and equivalent instructions consistent for every candidate?
Are protected topics, privilege assumptions, and unscored impressions excluded?
ChatGPT