# AI Customer Survey Report: Counts Before Conclusions

A survey can guide product decisions when readers understand who answered and what was asked. Use AI to organize a cleaned, authorized response set and draft explanations. Calculate or verify counts separately, keep skipped answers visible and distinguish a customer signal from a representative market finding.

## Your brief

1. Provide question wording, response options, survey dates and sample description. Remove names, emails and unnecessary respondent identifiers before processing.

   ____________________

2. Supply verified counts or a cleaned table with clear column meanings. State which questions allow multiple selections and which responses are missing.

   ____________________

3. Define the business decision and limits of the survey. Note recruitment bias, small samples or changed question wording that makes comparisons unreliable.

   ____________________

## Prompt

Draft a findings report from [survey questions and verified counts]. Include method, sample, per-question denominator, key themes, limitations and decision options. Do not invent percentages or claim statistical significance. Mark multi-select questions and skipped answers. Use authorized anonymized quotes only. Separate observed findings from recommendations and list calculations needing verification.

## Illustrative example

Illustrative survey: 30 customers responded, but only 24 answered the delivery question and 12 of those chose faster shipping. The report says 12 of 24 respondents to that question, not 50% of all customers. If multiple choices were allowed, totals may exceed the number of respondents; that condition belongs beside the chart.

## Acceptance checks

- [ ] Do counts, denominators, skipped answers and multi-select rules match the source?
- [ ] Are conclusions limited to the actual sample, with recruitment and wording limitations visible?
- [ ] Can every quote, theme and chart be traced to authorized data without identifying respondents?

Do not combine survey answers with unrelated analytics to imply a causal result. A respondent preference is not proof that a feature will increase sales; frame it as a hypothesis to test with an appropriate follow-up.

---

Brainbaby · https://brainbaby.ai/blog/ai-customer-survey-findings-report

CC BY 4.0 — https://creativecommons.org/licenses/by/4.0/