Content Writing · Long-form writing
The Best Prompt for Writing a Case Study
The best prompt for a case study treats your supplied facts as the only material allowed. This prompt has the model number every fact first, choose the result that matters most to the intended reader, and then write the challenge, solution and results with a fact reference after each claim. Quotes are used word for word or flagged as needed. You also get a list of gaps and questions to take back to the customer.
Customer Case Study Prompt
Drafts a customer case study strictly from the facts you supply, with every claim referenced and missing quotes or baselines listed.
- Works with
- ChatGPT, Claude, Gemini
- Difficulty
- Intermediate
- Complexity
- Multi-part reply
- Variables
- 7
- Version
- 1.0
Customize this prompt
Fill in what you know. Anything left blank stays as a [bracketed placeholder]. Nothing you type leaves your browser.
7 variables · fields marked * matter most
# Role You are a B2B content writer drafting a customer case study. The customer and a legal reviewer will both read it, so every statement has to trace back to something I gave you. # Facts supplied - Customer: [Customer] - Product or service they used: [Product or service used] - Who will read this case study: [Who will read the case study] <challenge> [The challenge] </challenge> <solution> [The solution] </solution> <results> [Results] </results> <quotes> [Customer quotes] </quotes> # Task 1. Fact inventory. Before writing, list every usable fact from the material above as a numbered item (F1, F2 and so on). For each result, note whether it has a number, a time period and a baseline. A result missing any of those is reported exactly as given, without rounding up or implying more. 2. Story line. In two sentences, state the change the customer went through in terms the reader above would care about. Choose the one result that matters most to that reader as the lead. 3. Draft the case study: - Headline: the customer and the main result, in plain words. - Summary: three or four lines covering the customer, the problem, what they did and the outcome. - The challenge: what was going wrong, what it cost them and what they had tried, as far as the facts say. - The solution: what was put in place, how it was rolled out and who used it. Describe the customer's decisions, not a tour of product features. - The results: each outcome with its number, period and baseline where supplied. - What is next, only if the facts mention it. 4. Put the fact number in brackets after each factual sentence, for example [F4], so a reviewer can check it. # Constraints - Supplied facts only. Do not add industry statistics, typical benefits or details about the customer's business that I did not give you. - Quotes are used word for word or not at all. Do not write, extend or polish a quote. Where a quote would strengthen a section and none exists, write [QUOTE NEEDED: who, about what]. - Do not turn a sequence into a cause. If a number improved after the rollout, say that; credit the product for the change only where the customer did. - The customer is the subject of the story. No superlatives, and no adjectives attached to results. # Output format 1. Fact inventory. 2. Story line. 3. The case study, 500 to 700 words, with fact references. 4. Gaps: [QUOTE NEEDED] markers, results missing a baseline or period, and up to five questions for the customer that would most improve the piece.
Use it when you are
- Turning customer interview notes and account data into a first case study draft
- Preparing a draft that the customer's team and legal can check line by line
- Finding out what to ask a customer before the interview, by running the prompt on what you already know
- Rewriting a feature-heavy success story so the customer is the subject
Why this prompt works
A fact inventory sets the boundary
Numbering every usable fact before drafting turns the instruction to use supplied facts only into something checkable. The model has a closed list to write from, and you can see at once whether it misread an input. Anything in the draft without a matching number stands out as a problem.
References make review fast
Case studies stall in approval because someone has to confirm each claim. A bracketed reference such as [F4] after each factual sentence lets the customer or a legal reviewer trace it to the source in seconds. Strip the references once the piece is approved.
Quotes are protected from improvement
Polishing or inventing a customer quote is the most damaging thing a draft can do, because it puts words in a real person's mouth. The word for word rule, with a [QUOTE NEEDED] marker as the alternative, keeps the model away from that and tells you which quotes to request.
Results are reported with their limits
A number without a period or a baseline says little, and a model will often smooth that over. Checking each result for all three, and forbidding claims of cause the customer did not make, keeps the results section accurate and produces useful follow-up questions.
The reader chooses the lead
The same project can be a cost story, a speed story or a staff story. Asking who will read the case study and making the model pick the result that matters most to that person gives the piece a focus that a generic challenge, solution, results template lacks.
How to use it
- Gather your interview notes, account data and approved quotes before you start. Record where each number came from.
- Describe the reader as a role at a prospective customer, not as a general audience.
- Paste facts under challenge, solution and results as they are; do not improve them on the way in.
- Check the fact inventory against your notes and correct anything misread before reading the draft.
- Send the gaps list and questions to the customer, add their answers to the inputs, and run the prompt again.
- Get written approval from the customer for the final text, then remove the fact references.
AI model compatibility
| Model | Status | Notes |
|---|---|---|
| ChatGPT | Expected to work | If a fact reference looks wrong, ask it to quote the source line for that reference; this is a quick way to catch a claim that drifted from the input. |
| Claude | Expected to work | Each input has its own XML-style tag, following Anthropic's guidance on separating supplied material. You can paste a full interview transcript inside the quotes tags. |
| Gemini | Expected to work | Keep the fact inventory step even for short inputs. Listing the facts first gives the draft a fixed set of material to draw from. |
“Expected to work” means the prompt uses plain, model-independent structure and has been reviewed for that — not that it has been run on that model. A model is marked Tested only when a test record is published below.
Worked example
Scenario: A fictional scheduling software company drafting a case study about a fictional dental group. Every detail is illustrative.
# Role You are a B2B content writer drafting a customer case study. The customer and a legal reviewer will both read it, so every statement has to trace back to something I gave you. # Facts supplied - Customer: Harbor Lane Dental, a group of six dental practices - Product or service they used: Chairside, an appointment scheduling and reminder platform - Who will read this case study: Operations managers at multi-location healthcare practices evaluating scheduling tools <challenge> Each practice kept its own calendar. Front-desk staff phoned patients the day before to confirm. No-shows were tracked on paper at four of six sites. Operations manager says cancellations were often not refilled. </challenge> <solution> Rolled out one practice at a time over ten weeks. Shared calendar across all six sites. Automated text reminders at 3 days and 1 day. Waitlist feature offers cancelled slots to patients who asked for earlier appointments. Front-desk leads trained in one session per site. </solution> <results> No-show rate at the two sites that tracked it digitally went from 11% to 6% comparing the quarter before rollout to the quarter after. Front-desk staff no longer make confirmation calls. Waitlist filled cancelled slots (no count available). </results> <quotes> Operations manager: 'The front desk got their mornings back. That was the first thing people noticed.' </quotes> # Task 1. Fact inventory. Before writing, list every usable fact from the material above as a numbered item (F1, F2 and so on). For each result, note whether it has a number, a time period and a baseline. A result missing any of those is reported exactly as given, without rounding up or implying more. 2. Story line. In two sentences, state the change the customer went through in terms the reader above would care about. Choose the one result that matters most to that reader as the lead. 3. Draft the case study: - Headline: the customer and the main result, in plain words. - Summary: three or four lines covering the customer, the problem, what they did and the outcome. - The challenge: what was going wrong, what it cost them and what they had tried, as far as the facts say. - The solution: what was put in place, how it was rolled out and who used it. Describe the customer's decisions, not a tour of product features. - The results: each outcome with its number, period and baseline where supplied. - What is next, only if the facts mention it. 4. Put the fact number in brackets after each factual sentence, for example [F4], so a reviewer can check it. # Constraints - Supplied facts only. Do not add industry statistics, typical benefits or details about the customer's business that I did not give you. - Quotes are used word for word or not at all. Do not write, extend or polish a quote. Where a quote would strengthen a section and none exists, write [QUOTE NEEDED: who, about what]. - Do not turn a sequence into a cause. If a number improved after the rollout, say that; credit the product for the change only where the customer did. - The customer is the subject of the story. No superlatives, and no adjectives attached to results. # Output format 1. Fact inventory. 2. Story line. 3. The case study, 500 to 700 words, with fact references. 4. Gaps: [QUOTE NEEDED] markers, results missing a baseline or period, and up to five questions for the customer that would most improve the piece.
What a good response should contain
- A numbered fact inventory that notes the no-show figure covers two of six sites and that the waitlist result has no number
- A lead built on the no-show reduction or on front-desk time, chosen with operations managers in mind
- The operations manager's quote reproduced exactly, with no additional quotes written
- The no-show result stated with its scope and comparison period, not generalized to the whole group
- Gaps that ask for a waitlist count, data from the other four sites and a quote about the rollout
This is an illustrative scenario showing how to fill in the prompt and what to look for. It is not a recorded model output. Recorded outputs will be published here once this prompt has a test record.
Testing and editorial review
Status: Untested. Reviewed for structure and clarity by the editorial team. No recorded model test yet.
No test record has been published for this prompt yet. When it is tested, this section will show the model and version, the date, the inputs, the actual output and notes against our seven-point rubric. We do not publish numerical scores.
| Author | TheBestPromptFor Editorial Team — The editorial team writes, reviews and maintains the prompts and guides on TheBestPromptFor. Prompts are drafted with AI assistance and reviewed against published editorial standards. |
|---|---|
| Published | November 15, 2026 |
| Last reviewed | November 15, 2026 |
| Last tested | Not yet tested |
| Version history |
|
Limitations and things to check
- The model cannot confirm that your facts are true or that the customer has approved them. Approval remains your responsibility.
- Fact references can occasionally point to the wrong item. Spot-check them, especially for numbers.
- Thin inputs produce a thin case study with a long gaps list. The prompt will not fill space with general claims.
- It produces text only. Design, pull quotes, logos and permission to use the customer's name are handled separately.
Alternative prompts
Content Writing
Blog Post Writing Prompt
Turns your own notes and experience into a focused blog post with one clear argument, and marks every spot where it needs a real example or a fact check.
- ChatGPT
- Claude
- Gemini
Content Writing
Conversion Copywriting Prompt
Works out what a page must make the reader believe, then writes the copy in that order using only the proof you supply.
- ChatGPT
- Claude
- Gemini
Research
User Interview Discussion Guide Prompt
Writes a timed discussion guide for customer or user interviews: screener, opening script, behavior-based questions with probes, and a table tying each question to a learning goal.
- ChatGPT
- Claude
- Gemini
E-commerce
Customer Review Analysis Prompt
Turns pasted reviews into counted themes with verbatim quotes, causes of complaints, customer language and recommendations graded by support.
- ChatGPT
- Claude
- Gemini
Related resources
Guide
How to Improve AI Responses
A diagnostic approach to weak AI answers: name the fault, trace it to the gap in your prompt, then follow up precisely or rewrite and start again.
Technical guide
Prompt Chaining: Breaking a Task Into Linked Prompts
How to split a large task into a sequence of focused prompts, design the handoffs between them, and debug the chain when a step goes wrong.
Frequently asked questions
What if I do not have hard numbers?
Supply what you have: what changed in the customer's daily work, what they stopped doing, what they say about it. The prompt will report those outcomes as given and list the missing figures under gaps. A specific qualitative result is more credible than a vague percentage.
Can the model write a placeholder quote for the customer to approve?
This prompt will not, by design. Drafted quotes tend to survive into published copy, and they read as marketing language. Use the [QUOTE NEEDED] markers to ask the customer a direct question on that subject, then paste in what they actually say.
How long should a case study be?
The prompt asks for 500 to 700 words, which suits a web page or a two-page PDF. If you need a one-paragraph version for a sales deck, ask for it as a follow-up using the same fact inventory so nothing new is introduced.
Sources and references
- Prompt engineering overview — Anthropic
- Prompt engineering guide — OpenAI
Was this prompt useful?
No account needed. Do not include personal information.