Raw Questions → Better Bank

Audit a raw question bank for bias, redundancy, missing coverage, and low-value prompts.

AI Pill #0231 min readDiscussion guidesChatGPTClaudeNotebookLM

This week's problem

Question banks grow by accumulation. Old questions stay, new questions get added, and nobody checks whether the bank is balanced, neutral, or still tied to the objective.

3 Useful AI Links

01ChatGPT

Classify, rewrite, and organize questions by objective, difficulty, and expected insight.

02Claude

Audit a long question set for duplication, ambiguity, bias, and logical flow.

03NotebookLM

Check coverage against an approved source set and identify unsupported or missing areas.

Workflow To Steal

Raw question bank + objectives
Tag each question by purpose
Remove duplicates and leading language
Find coverage gaps and weak distractors or probes
Publish a balanced, prioritized bank

Copy-paste prompt

Audit and rebuild this question bank so every question has a clear purpose.

Use case: [Advisory board, interview, workshop, assessment, or research]
Audience: [Audience]
Objectives: [What the bank must uncover or test]
Approved source material: [Paste or attach]
Question bank: [Paste]
Tag every question by objective and question type.
Flag duplicates, ambiguity, leading language, double-barrelled questions, and trivial items.
Rewrite weak questions and explain the change.
Identify objectives with too much or too little coverage.
Rank questions as Essential, Useful, or Remove.
Return the revised bank in a logical order.

Do not create false balance when the evidence is one-sided.

Do not add facts outside the approved source set.

Keep wording appropriate for the audience.

Best for

Discussion guidesAdvisory boardsKOL interviewsMarket researchInsight-gathering sessions

Verdict

A bigger question bank is usually a worse question bank. Use AI to cut, rebalance, and sharpen it before the session exposes the flaws.