Startup & Strategy

Verified against Claude · 2026-08-05

Turn a stack of customer discovery notes into patterns you can actually act on

Synthesizes raw notes from multiple customer discovery interviews into patterns that repeated at least twice, with verbatim quotes preserved — not one loud opinion mistaken for consensus.

ClaudeGeminiChatGPT3 fillable variables

The prompt

Ready to copy — highlighted parts are example details you can swap.

You are synthesizing customer discovery interviews. Your job is to find what actually repeated across interviews, not to elevate the most articulate or most recent answer into a conclusion the data doesn't support.

CONTEXT
Number of interviews: 9
Raw interview notes, labeled by interviewee if possible: Interview 1 (front-desk lead, single clinic): "Honestly the hardest part is just knowing who to see first when three people walk in at once"...
Interview 2 (front-desk lead, second clinic): "I don't trust an app to tell me who's worse off, that's a judgment call"...
Hypothesis being tested, if any: That front-desk staff want triage help, not a triage decision made for them

EXTRACTION
Extract the problem statements, current workarounds, and objections mentioned in each interview, keeping track of which interview each point came from — this tracking matters for the pattern rule below, so don't lose it in summarizing.

PATTERN RULE
Group points into a theme only when the same underlying point appears in at least 2 separate interviews. Anything said by only one interviewee gets labeled "Outlier — not yet a pattern," never promoted into a theme just because it was memorable, strongly worded, or came from a customer you particularly want to be right.

VERBATIM PRESERVATION
For each real theme, quote the closest-to-verbatim language interviewees actually used, not your paraphrase of it, so the exact words are available for messaging or copy work later — a founder's tidied-up summary of what a customer meant is not the same asset as the customer's own words.

HYPOTHESIS VERDICT
If That front-desk staff want triage help, not a triage decision made for them is given, state plainly whether the notes support it, contradict it, or leave it untested. Ambiguous or mixed evidence must be reported as ambiguous — never rounded up to "confirmed" because most of the signal pointed that way.

OUTPUT FORMAT
Themes (with interview count and verbatim quotes), Outliers, then a Hypothesis Verdict section if applicable.

Customize

Optional — swap in your own details for the highlighted parts above.

Why this works

The 'appears in at least 2 interviews before it counts as a theme' rule is a standard qualitative-research safeguard against single-loud-customer bias, where one especially articulate or emotionally intense interview gets remembered and repeated internally as 'what customers think,' even when nobody else in the sample said anything close to it. Preserving verbatim quotes instead of paraphrasing them keeps the customer's own words available for later use in landing pages and outreach copy, since messaging built from a customer's actual phrasing consistently tests better than a founder's tidied-up summary of what they think the customer meant — the difference between 'I don't trust an app to tell me who's worse off, that's a judgment call' and a paraphrase like 'staff want more control over triage decisions' is the difference between copy that lands and copy that sounds like every other SaaS pitch. Refusing to round ambiguous or mixed evidence up to 'confirmed' directly counters the confirmation bias every founder brings into discovery interviews for an idea they're personally invested in — it is exactly the failure mode this synthesis step exists to catch, since a founder reading their own raw notes has every incentive to read supportive intent into an ambiguous quote, while a synthesis pass held to an explicit ambiguity rule doesn't get to make that call quietly on the founder's behalf.

Verified against

Claude Sonnet 5 · 2026-08-05

Gemini 3 Pro · 2026-08-01

Changelog

  • 2026-08-05 Initial publish, verified against Claude Sonnet 5 and Gemini 3 Pro.

Need this built into your business?

If a prompt isn't enough — custom software, built and maintained for you — that's Scult's day job.

EXPLORE CUSTOM SOFTWARE
All Startup & Strategy prompts

Check your AI visibility

One URL in, a 0–100 score and the exact fixes out.

RUN THE CHECK

Browse all the tools

15 tools across six categories
13 of them never send your data anywhere

Free · No signup · No trial clock

SEE THE DIRECTORY