← All free guides
Free guide

The 5 Advisor Council Prompt for Claude Code

Ask one AI "is this a good idea?" and it'll usually find a way to say yes. This prompt runs 5 advisors who answer on their own, review each other with the names hidden, and give you 1 verdict.

A free guide by Zach Crawford

Here's the problem with asking AI about your business idea. It wants to help you. So if you ask "is this a good idea?", it'll usually find a way to say yes.

I don't ask one AI. I ask a council.

My own council runs 5 advisors on 4 different AI companies. They each answer on their own first, so nobody copies anybody. Then they review each other's answers with the names hidden. Then Claude checks the key facts against my real files and gives me one verdict.

This is a version you can run inside Claude, in one prompt. All 5 advisors run on Claude here, so each one gets a different job. That's what makes them disagree.

Step 1. Open Claude Code in a folder

The council needs Claude Code, because it starts 5 advisors at once. The regular Claude chat can't do that.

Make a folder for your idea. If you have any real facts about your business, put them in a file in that folder: your audience size, what you've sold, what your buyers have said, your prices. The council checks the verdict against them.

Open the Claude desktop app, switch from Chat to Code at the top, and pick that folder.

Step 2. Paste the council prompt

Fill in your idea at the top, then paste it in.

My idea: [your business idea in 1 to 3 sentences, like "a $297 course that teaches yoga teachers to sell private sessions"]

Run a council on my idea. Don't give me your own opinion first.

Step 1. Read every file in this folder for real facts about my business. List the facts you found in a few lines. If there are none, ask me 3 short questions first: who it's for, what they'd pay, and what proof I have that they want it. Wait for my answers.

Step 2. Start 5 advisors as separate agents, at the same time, so none of them sees the others' answers. Give each one my idea, the facts, and only its own job:
- The Buyer: you are the person this is for. Would you pay for it, at this price, right now? What would make you say no?
- The Skeptic: find the 3 biggest reasons this fails. Be specific. No pep talk.
- The Numbers: do the math. What has to be true for this to be worth it? Use only my facts. Mark every guess as a guess.
- The Competitor: who already solves this for these people, and why would they pick mine instead?
- The Builder: what is the smallest version I could sell this week, before I build the whole thing?
Each advisor answers in under 150 words and ends with one word: Build, Change or Drop.

Step 3. Hide the names. Label the 5 answers A to E in a random order. Start 5 reviewers as separate agents. Each one reads all 5 labeled answers and says which is strongest, which is weakest, and what all of them missed.

Step 4. You are the chair. Check every claim the advisors made against my facts. Throw out anything my facts don't back up, and say what you threw out.
Then give me one verdict:
- First line: Build it, Change it or Drop it.
- The 3 reasons that matter most.
- If it's Change it: exactly what to change.
- The one test I can run this week to find out if people will pay, before I build anything.

Rules: never make up numbers, quotes or results. If the advisors disagree, say so. The verdict is advice. I make the call.

It takes a few minutes. Claude will tell you as each advisor finishes.

Step 3. Read the verdict, then decide

Read the first line out loud. Then read the 3 reasons.

The council doesn't decide for you. It shows you what you missed. If it says Change it, the change is usually the cheapest thing you'll ever do for your business, because it happens before you build anything.

Run the test it gives you. Then run the council again with what you learned.

The limitations you should know

Want the next guide? Follow @iamzachcrawford