Case study prompt builder
A customer case study sells by proof, not adjectives. The best ones follow a simple arc — the challenge the client faced, the solution you delivered, and the results you can measure — and they lead with the strongest number. This builder takes your client, problem, solution, timeline, and outcomes and produces an LLM prompt that writes to that arc, foregrounds the hard metrics, and refuses to invent any figure you did not provide. You bring the prompt to your model and verify the draft.
How it works
You enter the client (named or anonymized), the problem they faced, what you did,
the implementation timeline, and the measurable outcomes. The tool assembles a
prompt that instructs the model to open with the headline result, structure the
body as challenge → solution → results, weave in your metrics, and write a
suggested pull-quote placeholder if you have no real quote. It bans fabricated
data and inserts [CONFIRM: …] placeholders for gaps. Everything runs locally
until you paste the prompt into your LLM.
What separates a persuasive case study from a weak one
The most common failure in case study writing is omitting the “before” state. Readers cannot judge an outcome without the baseline. Compare:
- Weak: “We helped Acme Corp improve customer retention.”
- Strong: “Acme Corp’s annual churn rate fell from 28% to 11% in six months.”
The second version gives a reviewer everything they need to evaluate the claim: a starting point, an ending point, a metric, and a time frame. The builder prompts you to supply all four, and the generated prompt instructs the model to foreground them in the opening line.
How to handle anonymized clients
B2B confidentiality often prevents naming clients. Effective anonymization preserves credibility without naming names:
- Use the industry and company size: “a Series B SaaS company with 120 employees.”
- Add the region if it matters: “a UK-based logistics provider.”
- Keep the metrics real — the numbers are what build trust, not the brand name.
The prompt builder adapts to either a named client or a descriptor. Anonymized case studies perform nearly as well as named ones when the outcome data is specific.
A note on attributed quotes
A quoted voice makes a case study more human and more believable. If you have a
real quote from your client, paste it into the notes — the prompt will weave it
in and mark it for attribution. If you do not have a quote yet, the prompt
inserts a clearly labeled [QUOTE: confirm with client before publishing]
placeholder, so you remember to collect one rather than publishing without it
or letting the model invent words in your client’s name.
Tips and examples
- Lead with the best number. If onboarding time dropped 60%, that is your opening line — the prompt is built to surface it.
- Quantify the before state too. “From 40 support tickets a day to 12” is stronger than just the after figure.
- Use a real quote if you have one. Paste it in; otherwise the prompt marks a placeholder rather than inventing words.
- Keep the timeline honest. Vague “rapid implementation” erodes trust. If it took three months, say three months — a realistic timeline is more credible.
- Verify every claim. A case study is a public, attributable document — check each figure against your records before it ships.