Daily Prompt Pack: Customer Discovery – Turn Five Raw Interviews Into A Decision Memo

I run discovery for a B2B tool where every call produces 40 minutes of rambling transcript. Here are the three prompts I use to get from that pile to a decision I can defend to my cofounder.

Today’s pack: customer discovery.

[Claude]

[Claude] Given these five interview transcripts pasted below, build a decision table with rows for each customer segment and columns: recurring pain, workaround they use today, money or time lost per month, and exact quote with participant ID. Then list the top 3 assumptions still unproven. Output as a markdown table plus a 5-bullet memo I can send my cofounder.

[ChatGPT]

[ChatGPT] Here are transcripts from five customer discovery calls. Extract every verbatim moment where the participant describes a workaround or manual fix. Group them by job-to-be-done, count how many participants mention each, and rank. Output a table with: job, workaround, participant count, and the strongest quote. Flag any workaround mentioned by only one person as anecdotal.

[Cursor]

[Cursor] In this repo, read the /interviews folder of markdown transcripts. Write a Python script that parses each file for pain statements, tags them by theme keyword list I define at the top, and outputs one CSV: participant_id, theme, quote, severity_score. Include a –dedupe flag that drops near-duplicate quotes. Then run it and show me the theme frequency counts.

Full archive at aicofounderstack.com/tag/prompt-pack.