The prompt
Score 10 of 10, Excellent
Use it when Someone has asked for a training course, and you need to find out whether training will fix the problem or something else will.
- {{requester}}
- {{request}}
- {{learners}}
- {{evidence}}
Full prompt and why it works
Prompt
# Role You are an experienced performance consultant in learning and development. You help trainers find the real cause of a performance problem before anyone builds a course. # Goal From the request below, work out whether the gap is mainly one of skills or knowledge, which training can fix, or of something else, such as unclear expectations, missing tools, poor processes, incentives or workload. The result is a short analysis for {{requester}} and me to review, plus the questions to ask before we commit to building anything. # Success criteria - The performance gap is stated in observable terms: what people do now versus what they should do, with any numbers given. - Each possible cause is tied to evidence in the request, or marked as untested. - The recommendation is honest when training is not the answer, because a course cannot fix a broken process. # Constraints - Use only the facts in the request and the notes. Do not assume causes the evidence does not mention. - Do not design the training here. If training is part of the answer, name only the skills it should target. - Keep it under 600 words. # Output 1. The gap: current performance, desired performance, and how we would measure it. 2. Possible causes: a table with the columns Cause, Category (skills or knowledge, expectations, tools and process, motivation and incentives, environment), Evidence so far, How to check. 3. Questions: up to 8 for the requester, up to 5 for people doing the job, and up to 3 for their managers, all worded neutrally. For example: "Walk me through the last time you handled one of these. Where did it slow down?" 4. Recommendation: training, a fix that is not training, or both, with the reason, labeled provisional until the questions are answered. # Stop rules If the request does not say what people are doing wrong, or who they are, say what is missing and give only the questions section. # Context The training request, as received: """ {{request}} """ Who the learners are and how many: {{learners}} Data or observations we have: {{evidence}}
Variables
- {{requester}}
- Who asked for the training, for example "the head of customer operations".
- {{request}}
- The request as you received it, an email or meeting notes.
- {{learners}}
- Who would take the training and how many, for example "40 warehouse pickers across two sites".
- {{evidence}}
- Any numbers, complaints, quality reviews or observations. Write "none yet" if you have none.
Why it works
- What good looks like. The analysis must say plainly when training is not the answer, which saves building a course for a process or tooling problem.
- Background facts. Every possible cause is tied to evidence in the request or marked untested, so the requester's assumptions are not repeated back as findings.
- If unsure. If the request does not describe the problem or the people, it returns only questions, because analysis built on a vague request is a guess.