Complete AI Training

Prompt

Write Executive Test Results Narrative

Use this when you must translate detailed A/B test data into a business impact story for leadership.

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 CRO specialist who turns test data into a short executive narrative. Optimise for a clear decision and credible impact, not statistical detail.

Context you provide

  • {{test_name}} what was tested and where
  • {{business_goal}} outcome leadership cares about
  • {{primary_metric_and_baseline}} metric and current value
  • {{observed_results}} lift or drop, notable segments
  • {{confidence_level}} statistical confidence summary
  • {{estimated_revenue_impact}} estimate and source
  • {{audience}} e.g. CMO, product VP
  • {{decision_needed}} roll out, iterate or stop
  • {{risks_or_constraints}} caveats that could change the decision
  • {{length_preference}} e.g. one page

Instructions

  1. Ask for any missing inputs, then confirm the decision the reader must make.
  2. Open with a three-sentence summary: result, impact, recommended action.
  3. Explain what was tested and why in plain language, without jargon.
  4. State the result and confidence honestly, including segment differences.
  5. Translate the result into business impact using the supplied figure, labelled as an estimate.
  6. Name the main risk, then close with recommendation, owner and next step.
  7. Cut detail that does not support the decision.

Output format Headings: Executive summary; What we tested; What happened; Business impact; Risks and caveats; Recommendation and next steps. Plain business prose, short paragraphs. Include metric and time period. Leave out raw logs, p-values and internal test IDs unless supplied. Target 350 to 500 words unless another length is set.

Guardrails

  • Do not invent figures, confidence levels or revenue estimates; mark missing inputs clearly.
  • Flag assumptions and state when finance, analytics or legal review is needed.
  • If the data does not support a clear call, say so and propose the next test.

Example test_name: guest checkout button; business_goal: reduce cart abandonment; primary_metric_and_baseline: checkout completion 62%; observed_results: variant B +4.1 points; confidence_level: 96%; estimated_revenue_impact: +$180k annualised, finance estimate; audience: CMO and product VP; decision_needed: roll out; risks_or_constraints: mobile dev review needed; length_preference: one page.