AI Personal Board of Directors
Build a researched panel of AI personas to examine any hard problem
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 6
- Confidence
- 98%
This method creates a simulated board of directors whose members represent different entrepreneurial, creative, financial, branding, or leadership perspectives. First, select people with complementary strengths rather than choosing several near-identical thinkers. Ask an LLM to construct the initial prompt, then research the distinctive experiences, methods, and qualities of every selected adviser. Incorporate those summaries so the model has more than a famous name to imitate. Present the board with a clearly stated problem and, when possible, supporting material such as a memo, deck, transcript, or recording. The model then produces separate responses through each perspective. The output is not a substitute for judgment; its value lies in exposing assumptions, questions, risks, and opportunities that the user might otherwise overlook.
Origin
Extracted from Marketing Against The Grain, where the host demonstrates building a free AI board featuring Steve Jobs, James Dyson, Sam Zell, Estée Lauder, Bernard Arnault, Oprah Winfrey, Charlie Munger, and Ralph Lauren.
Core principles
- 01Diverse experience reveals angles one thinker may miss
- 02AI can help design and improve its own prompt
- 03Research makes simulated perspectives more distinctive
- 04More context produces more relevant advice
- 05The user retains responsibility for every decision
How to run it
- 1
Select complementary advisers
Choose people whose experiences and strengths cover different dimensions of the problem. Include only personas that can contribute a meaningfully distinct lens.
Pro tip Mix operators, creators, investors, brand builders, and communicators rather than selecting only people from one discipline.
Watch out A board of similar thinkers will create repetition instead of useful cognitive diversity.
- 2
Ask AI to design the prompt
Tell the LLM that you want a reusable personal-board prompt and identify the selected members. Ask it to define how each member should analyze and respond to a problem.
Pro tip Use the model as a prompt-writing collaborator instead of trying to perfect the first prompt manually.
Watch out Do not treat the initial generated prompt as finished.
- 3
Research each perspective
Generate researched summaries focused on the unique experience and qualities that distinguish each adviser. Prefer source-supported material when research features are available.
Pro tip Research decision habits, operating principles, signature strengths, and recurring questions rather than collecting generic biographies.
Watch out A famous name without substantive context encourages shallow stereotypes.
- 4
Enrich the board prompt
Insert the research summaries into the reusable prompt and specify a consistent response structure. Leave a clear location for the user's idea, writing, or challenge.
Pro tip Preserve separate sections for each adviser so their perspectives remain distinguishable.
Watch out An excessively vague response format can cause the personas to collapse into generic advice.
- 5
Supply the problem context
State the challenge, desired result, constraints, and relevant evidence. Add a memo, deck, transcript, or recording when the problem warrants deeper analysis.
Pro tip Explain what has already been attempted and where you feel uncertain or stuck.
Watch out A basic input will usually produce high-level rather than tactical guidance.
- 6
Convene the board
Run the prompt and ask every adviser to respond independently from the researched perspective. Review where their recommendations agree, conflict, or expose a new question.
Pro tip Follow up with the most relevant persona for a tactical deep dive.
Watch out The generated response is a simulation, not an authenticated statement from the named person.
In the wild
The host asks the simulated board how Marketing Against The Grain could grow views and subscribers faster on YouTube. The personas produce different lenses: Steve Jobs emphasizes an intuitive visual and narrative experience, James Dyson emphasizes relentless testing, Sam Zell looks for an overlooked niche, and Oprah Winfrey emphasizes authentic audience connection.
→ One broad growth problem is reframed as several concrete areas for investigation and experimentation.
The host asks how AI will disrupt marketing, where to focus, and where the largest opportunities may appear. The board surfaces experience design, process efficiency, undervalued innovation, multidisciplinary risk analysis, and timeless brand building as distinct strategic lenses.
→ The user receives a wider opportunity map than a single undifferentiated answer would provide.
Common mistakes
Using names without research
A persona based only on a famous name is likely to produce clichés. Add researched context about the person's distinctive experience, qualities, and decision patterns.
Providing too little context
A one-sentence problem can demonstrate the method, but it limits the specificity of the advice. Use supporting documents and constraints for consequential decisions.
Treating simulation as authority
The board generates perspectives rather than verified opinions or final answers. The user must evaluate the recommendations and own the decision.
Is it for you?
Best for
It is best for founders, leaders, creators, and professionals confronting ambiguous problems without access to a formal advisory board.
Not ideal for
It is not ideal for decisions requiring verified facts, licensed professional judgment, or accurate predictions of what a real person would say.
From the transcript
“one of the best things to do is ask AI to help you write a prompt.”
“I want summaries that focus on the very unique experience and qualities that make each person great.”
“And then I said, Hey, can you add all this deep research into that prompt?”
From the episode
How to Think Like Steve Jobs in 10 Minutes [FREE Prompt]