Create your own board of directors in ChatGPT and finish decisions you've been hesitating on within the same day.
That heavy decision—if you call Masayoshi Son and Jeff Bezos into ChatGPT and let them debate, it'll be done today.
Should you raise prices? Hire someone? Pull out of that business?
There are decisions you've been holding onto for days alone, right?
To be clear, this isn't about writing text or summarizing minutes. Everyone knows that. This is about a way 90% of people aren't using ChatGPT yet: hiring 5 master managers with different perspectives simultaneously to debate your decisions.
The more important the decision, the fewer people there are to consult
For light consultations, there are plenty of people who will help.
But for truly heavy decisions, there aren't many people you can consult with as equals.
You can't show weakness to employees. Family doesn't understand the business details. If you ask a friend about raising prices, they just say, "Sounds fine."
So you carry it alone.
You go around in circles for days, write it down, erase it. Even after deciding, you wonder, "Was this really the right choice?"
But if you call Masayoshi Son, Kazuo Inamori, and Elon Musk into ChatGPT to debate, a half-day of circling moves forward in about 30 minutes.
It might sound like an exaggeration, but those who are doing it are doing it quietly.
ChatGPT can be a 'Board of Directors' rather than a 'Single Assistant'
Most people use ChatGPT as a Q&A assistant.
Ask a question. Get an answer. Done.
But the real way to use it is to run multiple personalities with different perspectives at the same time.
Overseas, this is already a standard technique known as a "virtual board of advisors."
The MIT Sloan Management Review has introduced examples of creating virtual boards based on historical leaders to draw advice on strategy, innovation, and ethics.
The author says: This is not to replace human advisors, but to broaden the perspective of judgment.
This is important, so I'll share it first.
You aren't trying to "reproduce the person's actual speech."
Instead, you have ChatGPT play that person's "thinking lens."
Imagine running several views at once: "How would Masayoshi Son see this situation?" or "What would Inamori-san get stuck on?"
Don't assume the person actually said this. You are just borrowing their thinking model. If you miss this point, it just becomes misinformation.
Why debating with 5 master managers improves decisions
The reason is simple.
Thinking with one head biases your perspective.
People can only see things through the lenses they are good at. Aggressive people only see the attack. Defensive people only see the defense.
That creates blind spots.
But if you intentionally clash different lenses, counterarguments and oversights that you would never have come up with yourself are forced to the surface.
For example, these five lenses:
- The Aggressive Investor: Someone who says to bet big no matter what.
- The Long-term Strategist: Someone who thinks about what happens in 3 years.
- The Harsh Skeptic: The so-called Devil's Advocate. Someone whose job is to find flaws.
- The Wise Humanist: Someone who speaks of people and trust, not numbers.
- The Practical Operator: Someone who thinks about tomorrow's reality and immediate effort.
When you hit these five, your judgment becomes three-dimensional.
And the decisive factor is that with AI, you can hire these lenses—as many as you want, as often as you want—for free.
With human advisors, you can't let them conflict this ruthlessly.
Choosing members (Who and Why)
A common mistake here is choosing based on "fame."
Instead, choose by lens.
Decide the roles first, then assign famous managers everyone knows. This order is important.
Suppose you're hesitating about a price hike.
First, decide the roles: A lens for bold investment judgment. A lens for profit and the front lines. A lens for working backward from the customer. A lens for thorough flaw-finding.
Then assign names: Masayoshi Son for the attack. Kazuo Inamori for profit and the front lines. Jeff Bezos for customer-centricity. For the flaw-finder, choose a harsh skeptic.
Why these people? Because everyone can imagine, "This person would probably think this way."
You, the reader, should swap them out to fit your theme.
For hiring decisions, increase the human-centric lenses. For exit decisions, add lenses strong in cutting losses.
Choose by lens. If you remember this, you can recombine members indefinitely.
Steps to 'Install' the Board in ChatGPT
Here is the actual method. 3 steps.
First: Define each member one by one.
Like writing on paper, decide the role, tone, and most prioritized values in bullet points.
For example:
Masayoshi Son role: Aggressive investor. Tone: Bold and quick. Priority: Whether the market can be taken.
Kazuo Inamori role: Guardian of profit and the front lines. Tone: Quiet and strict. Priority: Whether employees can work happily and profit remains.
Prepare this for the number of people you want to call.
Second: Create an instruction to fix the roles.
The framework looks like this:
"You are now my board of directors. The members are the following 5 people. Each member must speak from their respective standpoint. And please properly argue against each other's opinions. Finally, summarize one conclusion based on the discussion."
Paste the 5 definitions you just made after this.
Third: Save this so you can call it every time.
It's convenient to put it in ChatGPT's Projects feature or Custom Instructions.
Once made, you just throw in the consultation next time. You don't have to write from scratch every time.
Now, one personal board of directors is complete.
Actually having them debate and reach a conclusion
This is the step that differs most from a simple Q&A.
The trick is not to ask for a conclusion immediately.
First, let each member give an opinion. "Give me the opinions of these 5 people regarding this price hike."
Next, make them conflict. Just say, "Argue against each other's opinions."
This suddenly moves the discussion. The Inamori role might snap at the Masayoshi Son role, saying "The front lines won't hold."
And for the skeptic role, give them a clear job: "As the Devil's Advocate, tell me the biggest concern regarding this decision."
Here, the weaknesses you were trying not to see come out. It usually makes you flinch.
Finally, force a conclusion.
"Based on the discussion so far, narrow down what I should decide today to one thing."
Finish with this.
If you think "Wait, that premise is wrong" in the middle, don't hesitate to jump in. "Try again on the premise that we don't have enough staff."
You can redo the debate as many times as you want. You can't do that with a human board.
Scenarios where it works
Here are some scenes where this is effective:
Deciding on a price hike.
Before: You're alone with a calculator, only seeing faces of customers leaving, and you stop.
After: Attack, defense, and customer perspectives clash for 30 minutes, giving you a guideline for what price makes sense.
First full-time hire.
Before: You can't decide for weeks between anxiety over labor costs and the labor shortage.
After: The front-line lens and the long-term lens debate, providing the materials to judge if it's now or in six months.
Withdrawing from a loss-making business.
Before: Feelings of waste come first, and you drag it out.
After: A lens strong in cutting losses ruthlessly points out flaws, verbalizing the conditions to continue vs. close.
Allocating time to a new business.
Before: Only the desire to do it spins its wheels.
After: The aggressive investor role pushes you, the skeptic role brakes, and a perfect first step becomes visible.
What they all have in common is the gap between worrying alone for days with no one to talk to, and a state where a conclusion is reached through multi-faceted debate.
Of course, the output depends on the consultation, but it definitely moves you forward more than just spinning it in your head.
Don't follow blindly. Why it's still worth having.
Finally, an important warning.
These are not the person's actual words. They are thinking lenses played by ChatGPT. Never confuse the two.
The final judgment is always yours. The board won't decide for you. They only give you perspectives.
And for specific numbers or facts that come up in the debate, verify them separately. AI will confidently put out plausible-sounding numbers.
As long as you follow these three points, there is almost no risk.
Beyond that, the lesson I want to convey is this:
The quality of decision-making is determined by the number of lenses you hit it with.
If you decide only inside one head, blind spots never disappear. You can't see what you can't see.
So, don't make decisions alone. Intentionally pass them through a system of conflict.
With AI, you can hire advisors with different standpoints—as many as you want, for free, as often as you want. This was impossible just a few years ago.
Take just one decision you're currently holding and throw it to this board.
The time you spent spinning it alone will start to feel a little silly.
