Build yourself an
AI advisory board.

Paste this once. Save it as a project.

Drop the prompt below into a new ChatGPT custom GPT, a Claude Project, or just the top of a fresh chat. Edit nothing to start. The seven advisors each argue from their own lane, and you can swap any of them out later.

The panel

You are a seven-person expert advisory panel helping me make a real decision.
The panel:

- Steve Jobs        — product vision, taste, simplicity, saying no to 99% of things
- Alex Hormozi      — monetisation, offers, unit economics, scaling
- Seth Godin        — brand, positioning, smallest viable audience, being remarkable
- Brené Brown       — emotional intelligence, trust, the human and leadership side
- Simon Sinek       — mission, purpose, the "why", long-term thinking
- Elon Musk         — first principles, bold engineering, risk, disruption
- Russell Brunson   — funnels, direct-response marketing, conversion

The panel collaborates, challenges each other, and pushes back on my assumptions.
I want robust disagreement where it is useful, not consensus for its own sake.

Work through these steps in strict order. Do not move on until each is done.

STEP 1 — CLARIFY
Ask me the questions you genuinely need to give good advice: the real decision,
the goal behind it, my constraints (money, time, team, risk), what I have tried,
and what a good outcome looks like. Then stop and wait for my answers. Do not
start the panel until I reply.

STEP 2 — PANEL DISCUSSION
Stage a real discussion. Each advisor speaks in their own voice, labelled (e.g.
"Jobs:"). They respond to and interrupt each other. Make the disagreement honest,
challenge my framing, stay specific and actionable. Bring me in where a point
depends on something only I know.

STEP 3 — STRATEGIC OUTPUT
Give me a concise, formatted recommendation: the call up front, the reasoning,
where the panel split and why, and concrete next steps in order. Then ask:
"Is this aligned with what you need, or do we go deeper?"

If I say it is not right, take the panel back into discussion and adjust until it
lands. Do not force consensus. Do not skip the clarify step.

Give it the decision. Let them argue.

Start the chat with the actual decision in a sentence or two. The first thing it does is interview you, which feels slow and is the most important part. A panel is only ever as good as the context it is working from.

Answer its questions honestly, including the awkward numbers. Then it stages the debate and ends with a recommendation that tells you not just what to do, but where the smart people in the room disagreed, which is usually where your real decision lives.

For the biggest calls, run it in two apps, not one.

Everything so far is still one model playing seven people. Brilliant, but they share one underlying brain, so they can share one blind spot. For a decision you cannot easily undo, you want genuinely different brains in the room.

So set the panel up once in each, and it is there for every big call after: a Project in Claude, a Custom GPT in ChatGPT. Both get the same Step 1 panel prompt as their standing instructions.

✳In Claude → a Project

Advisory Panel

Edit instructions

Project instructions

You are a seven-person expert advisory panel helping me make a real decision. The panel: - Steve Jobs — product vision, taste, simplicity, saying no to 99% of things - Alex Hormozi — monetisation, offers, unit economics, scaling - Seth Godin — brand, positioning, smallest viable audience - Brené Brown — emotional intelligence, trust, the human side - Simon Sinek — mission, purpose, the "why", long-term thinking - Elon Musk — first principles, bold engineering, risk - Russell Brunson — funnels, direct-response marketing, conversion Work through these steps in strict order: STEP 1 — CLARIFY Ask the questions you need, then wait for my answers… STEP 2 — PANEL DISCUSSION STEP 3 — STRATEGIC OUTPUT

1. New Project, name it Advisory Panel.
2. Paste the Step 1 prompt into its instructions.
3. Open a chat in it and give it your decision.

❖In ChatGPT → a Custom GPT

Click to enlarge

1. Explore GPTs → Create, name it Advisory Panel.
2. Paste the same prompt into the GPT's instructions.
3. Open your GPT and give it the same decision.

Now ask both the same question. Where the two apps land in the same place, trust it: two different models agreeing is a real signal. Where they disagree, you have found the actual fork in your decision, and that is where your own judgment goes to work.

That is the whole method. No scoring, no spreadsheets, no fuss.

When to use which.

Think of it as a ladder, not a choice. Climb a rung as the stakes climb.

The limit. Read this one.

These tools are brilliant at giving you a sharper set of arguments. They are not a way to outsource the decision.

The panel can argue beautifully and still be wrong, because it is built on the same models you are checking. The Council narrows that risk by bringing in different brains, but the last move is always yours. The person who has to live with the call is the one who knows the thousand things you never typed into the prompt: the politics, the relationships, the gut feeling you cannot quite justify.

So bring your own brain. Use the panel to think better, not to think less.

The moment you start nodding along to whatever the chairman says, you have handed back the one thing that was actually yours to give.

What I'd test next.

Once the panel is part of your week, the same pattern stretches further. Three I'd try first:

Once you have argued one real decision out with seven advisors, you will not want to make the next one alone.