Use when you have raw customer interview material and need a case study a prospect would believe.
case-study-from-interview-notes.md
You are drafting a customer case study from interview material. You are not writing testimonial copy and you are not inventing outcomes.
Customer, what they bought and when: {{CUSTOMER_CONTEXT}}
Interview notes and quotes: {{INTERVIEW_NOTES}}
Figures the customer has agreed we can publish: {{APPROVED_METRICS}}
Who reads this and what they are deciding: {{READER_DECISION}}
Output:
1. Claim ledger table: Claim | Source in {{INTERVIEW_NOTES}} or {{APPROVED_METRICS}} | Attributed to them or asserted by us | Needs their sign-off (yes/no).
2. The case study in five parts: the situation before, what they tried first, what changed, what happened, and what they would tell someone facing the decision in {{READER_DECISION}}. Ground the first part in {{CUSTOMER_CONTEXT}} alone.
3. Two opening quote options, verbatim.
4. Sign-off list: every ledger row marked yes, grouped by who has to approve it.
Rules:
- Every figure must appear in {{APPROVED_METRICS}}. Do not calculate percentages from numbers that were not published together.
- Where the notes are vague on cause, report what the customer said happened, not why it worked.
- Do not tidy a quote into marketing language. Mark any shortened quote [lightly edited].
- If the interview never establishes a before state, say so. Without one this is a testimonial, not a case study.
- No em dashes. British spelling.
Replace each placeholder with your own detail. The more specific you are, the less the model invents.
When should I use this rather than a general writing prompt?
When a prospect has to believe the result. A general prompt fills the gaps with confident outcome language and calculates percentages nobody agreed to. This builds the claim ledger first and is told to say plainly when the interview never established a before state, which is when you have a testimonial rather than a case study.
What has to be agreed before I run it?
The customer context including what they bought and when, the interview notes and quotes, the figures they have agreed you can publish, and who reads this and what they are deciding. Only approved metrics go in, and percentages are not derived from numbers that were never published together.
What does it produce for the approval process?
A claim ledger marking source, attribution and whether sign-off is needed, the study in five parts from situation to advice, two verbatim opening quote options, and a sign-off list grouped by approver. Send the ledger for approval rather than the draft. Customers return a ledger far faster than prose.
What cannot be worked around?
Writing round a missing before state. Without it there is nothing to attribute the change to, and the piece cannot survive a prospect asking what improved. Book the second call. File the ledger with the published page as well, for the day someone queries a figure two years later.