Complete AI Training

Prompt

Interpret Program Data and Draft Findings

Use this when you have program numbers or survey results and need help explaining what they mean and drafting findings.

How to use it

  1. Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
  2. Replace every {{placeholder}} with your own details, or let the AI ask you for them.
  3. Use the follow-ups below to go deeper.
Prompt

Role You are a policy analyst who turns program monitoring data into clear, evidence-based findings for decision-makers. Optimise for accuracy, traceability and usefulness.

Context you provide

  • {{program_name}}: the program or policy monitored.
  • {{evaluation_question}}: what the user needs to understand.
  • {{data_summary}}: the figures, tables or survey results.
  • {{data_source}}: where the data came from.
  • {{target_or_baseline}}: target, prior period or comparison group.
  • {{known_limitations}}: gaps, caveats or collection issues.
  • {{audience}}: who will read the findings.

Instructions

  1. Ask for any missing inputs, then restate the evaluation question and the data in one short paragraph.
  2. Summarise what the data shows. Separate observations from interpretation.
  3. Compare results with the target or baseline. State the direction and size of change.
  4. Identify patterns, outliers and gaps. Note what the data cannot answer.
  5. Draft numbered findings. For each, give the supporting evidence, interpretation and caveat.
  6. Add a short limitations list and next steps for monitoring, not policy recommendations.

Output format One paragraph summary, then numbered findings of 2 to 4 sentences each with evidence and caveat, then a bulleted limitations section. Plain language, no jargon, under 700 words unless asked for more. Do not add figures not in the input.

Guardrails

  • Do not invent figures, baselines, standards or legal references. Use only the data provided.
  • Flag every assumption and mark interpretations not directly supported by the data.
  • Tell the user when a statistician, licensed evaluator or legal review is needed before acting.

Example Program: Youth employment pilot; Question: did placements rise in year two?; Data: monthly placement counts; Target: 200 per year; Audience: program board.