The ChatGPT Playbook

Most people use ChatGPT.
Few people know how to use it.

Seven case studies structured on the Harvard Business School Case Method. One system. Change how you ask. Permanently.

7
Case Studies
4
Core Techniques
50
Power Prompts
Free preview below

Three questions. Be honest.

Find out where you actually stand. Think about the last time you used ChatGPT and answer honestly.

Question 01
Did you assign ChatGPT a specific expert role: a named professional with a domain, not just "be helpful"?
Question 02
Did you tell it who the output is for and what would make it fail or get rejected?
Question 03
Did you use the first response as-is, without rewriting, regenerating, or significantly editing it?

Four variables. Every prompt. Every time.

Every output gap traces back to one or more of these. Miss one and the model fills it with its default assumption. That assumption is almost never right.

01
Role
A named expert. Not just "be helpful." The model performs to the role you assign.
02
Context
Audience, stakes, constraints. What the model needs before the task.
03
Task
One precise directive. One outcome. No ambiguity.
04
Format
Length, structure, tone, what to leave out.

Now let's put it to work.

The 9am Deadline

A real scenario, six steps: situation, the gap, the technique, the prompt, the output, and why it works. Structured on the Harvard Business School Case Method.

Situation
Situation Challenge Technique Prompt Output Analysis
Step 1 of 6 ยท Situation

The scenario

What happened
Your manager forwarded a 40-page market analysis report at 7pm. She needs a 1-page executive summary ready for the leadership team's budget meeting tomorrow morning.
The audience
The leadership team. Time-poor, data-literate, and using this summary to approve or reject a โ‚ฌ2M budget allocation. If the summary is vague, so is their decision.
The stakes One hour. One page. A budget decision that depends on it. The quality of your summary will be read as a proxy for your judgment.
1 of 6
Step 2 of 6 ยท The Challenge

What most people do

Your prompt to ChatGPT
"Summarise this report for my leadership team."

Twelve words. No role. No audience context. No decision stakes. No format constraint. The model has nothing to work with except the document itself.

ChatGPT response
This report provides a comprehensive overview of the current market landscape. Key findings include significant growth in the digital segment (up 23% YoY), increasing competitive pressure from emerging players, and shifting consumer preferences toward mobile-first experiences. The report recommends a strategic review of current positioning. In conclusion, the market presents both challenges and opportunities that require careful consideration going forward.
Generic. Flat. Unusable.
2 of 6
Step 3 of 6 ยท The Technique

Build the prompt

Tap each variable to load it. Four variables. One structured prompt.

01 ยท Role
Who does ChatGPT become?
Senior management consultant specialising in executive communication.
02 ยท Context
What does it need to know first?
Leadership team, time-poor, making a โ‚ฌ2M budget allocation decision from this summary.
03 ยท Task
What exactly should it produce?
Three key findings, the main risk, and the recommended next step.
04 ยท Format
What should it look like?
Max 250 words. No jargon. Lead with the conclusion, not the methodology.
3 of 6
Step 4 of 6 ยท The Prompt

All four variables, assembled

Each colour marks one variable. The structure is always the same. The content changes.

You are a senior management consultant specialising in executive communication. My leadership team needs a 1-page summary of a 40-page market analysis report. They are time-poor and will use this summary to make a budget allocation decision of โ‚ฌ2M. Identify the three most decision-relevant findings, the single most important risk, and the recommended next step. Max 250 words. No jargon. Lead with the conclusion, not the methodology.
Role
Context
Task
Format
4 of 6
Step 5 of 6 ยท Expected Output

The difference

ChatGPT: generic prompt
This report provides a comprehensive overview of the current market landscape. Key findings include significant growth in the digital segment (up 23% YoY), increasing competitive pressure from emerging players, and shifting consumer preferences toward mobile-first experiences. The report recommends a strategic review of current positioning. In conclusion, the market presents both challenges and opportunities that require careful consideration going forward.
No recommendation. No decision logic.
ChatGPT: structured prompt
Recommendation: Allocate the โ‚ฌ2M to the digital segment.Three findings support this: (1) Digital grew 23% YoY while traditional declined 8% โ€” the shift is structural, not cyclical. (2) Two new entrants captured 15% market share in 18 months using mobile-first models we do not yet operate. (3) Customer acquisition cost in digital is 40% lower than our current channel mix.

Main risk: The window is narrow. Both competitors are closing Series B rounds this quarter โ€” 6-9 months before they can scale.

Next step: Approve the digital infrastructure investment in this meeting so procurement can begin Q3 onboarding.
Decision-focused. Structured. Ready to use.
5 of 6
Step 6 of 6 ยท Why It Works

The analysis

1
Role set the standard."Senior management consultant specialising in executive communication" told the model to write for busy decision-makers โ€” not to explain the report, but to advise on it.
2
Context defined what "relevant" means.The model knew a โ‚ฌ2M allocation was at stake. That single fact shaped every choice โ€” which findings matter, what risk to flag, what the recommended action is.
3
Format eliminated the most common failure."Lead with the conclusion, not the methodology" is the single highest-impact format instruction for executive audiences. Without it, most models bury the recommendation in paragraph three.
You now know the system
You have a situation like this coming up this week.
Six more case studies. Every prompt type you actually face at work and in life. Apply the system, or keep guessing.
Get the full playbook for $19
6 of 6
"I've been using ChatGPT for months and thought I was doing fine. Went through the case study, tried the structured prompt on an actual client brief โ€” the difference was immediate. It's not about the tool, it's about the structure. This made that obvious in 10 minutes."
Claudio
Senior Data Analyst ยท data-works.ch
What you unlock
The complete ChatGPT Playbook
Case Study 01 ยท Previewed above
The 9am Deadline
Your manager sent a 40-page report at 7pm. Leadership meeting tomorrow. You have one hour to produce something they'll actually use.
Case Study 02
The Salary Conversation
You know you're underpaid. You just don't know how to ask without making it awkward โ€” or worse, damaging what you have.
Case Study 03
The Email Sitting in Drafts
The one you haven't sent because the tone isn't right yet. Boss, client, landlord, colleague: the prompt that finally gets it right.
Case Study 04
The LinkedIn Post That Doesn't Sound Like AI
You have something worth saying. You just don't want to sound like everyone else doing thought leadership with the same tool.
Case Study 05
The Job Application
A cover letter that sounds like you โ€” not a template every other applicant is generating with the exact same prompt.
Case Study 06
The Refund You Deserve
Complaint to an airline, landlord, or service provider that actually gets a response โ€” firm, specific, impossible to ignore.
Case Study 07
The Document You Don't Understand
A clause in your lease. A medical report. A financial statement. Explained in plain language โ€” fast, accurate, no guessing.
The Edge Cases
3 Failure Modes โ€” and the Fix
What triggers vague outputs, hallucinations, and wrong register โ€” and the exact recovery prompt for each.
The Mastery Stack
Advanced Techniques
Prompt Chaining, Output Validation, Parallel Generation โ€” with worked examples.
Bonus
50 Power Prompts โ€” ready to use
Fully written prompts across Writing & Communication, Analysis & Decision-Making, Learning & Explanation, and Career & Professional. Copy, paste, adapt.
You've seen the architecture

The people ahead of you
aren't smarter.
They stopped improvising.

Seven case studies. Three failure modes with recovery prompts. 50 production-ready prompts. The Mastery Stack. The complete framework โ€” applied to your next real situation, not a hypothetical.

$19
One-time ยท Instant access ยท No subscription
Unlock the full playbook โ†’
30-day money-back guarantee. No questions asked.
Secure checkout via Lemon Squeezy