Build Your Own AI Council: A Business Decision-Making Prompt Inspired by LLM Council

There is a quiet assumption baked into most conversations about AI in business: that the technology’s job is to give you answers faster. Type a question, get a response, move on.

At Lumen, we think that framing misses the point.

The real value of AI is not speed of answers. It is quality of thinking.

The best business decisions rarely come from a single confident voice. They come from tension: from being challenged, stress-tested, and forced to defend your reasoning before you commit resources, reputation, or client trust.

That is the culture we build at Lumen. AI does not replace it. Used well, it amplifies it.

So here is a tool we use — adapted from the wider LLM Council idea — and one you can apply inside your own business.

The Problem With How Most People Use AI

Most people use AI like a vending machine. Insert prompt, receive answer, treat it as truth.

The danger is not only that the answer may be wrong. The bigger danger is that the answer may be plausible. A well-written, confident, single-perspective response is easy to accept, especially when you are already leaning toward a decision and looking for validation.

Good consulting culture is built on the opposite instinct.

Before we put a recommendation in front of a client, it gets pulled apart internally. Someone plays devil’s advocate. Someone asks whether we are solving the right problem. Someone spots the angle nobody else saw because they were not buried in the detail. Someone asks what happens if the client, the market, or the numbers do not behave the way we expect.

That is not negativity. It is responsibility.

The insight is simple: you can ask AI to help create that same disciplined challenge.

Credit and Inspiration

The AI Council approach in this article is adapted from the wider LLM Council idea, including Andrej Karpathy’s LLM Council project.

Karpathy’s project uses multiple AI models to examine the same question, review and rank each other’s responses, and produce a final synthesis through a Chairman role.

The version used here adapts that council-style thinking into a practical business decision-making prompt.

Lumen’s contribution is not the invention of the council concept. Lumen’s contribution is applying this method to business culture: better challenge, better listening, sharper thinking, faster testing, and stronger accountability in the age of AI.

The AI Council Prompt

Instead of asking AI for one answer, ask it to convene a council.

Five distinct advisors. Five different mandates. One final decision.

Paste this prompt into your AI tool, add your idea at the bottom, and include the goal, context, constraints, target customer, risks, budget, timeline, and what you have already considered.

You are my AI Council.

I will give you an idea, question, or business decision.

Five advisors must respond separately:

1. The Contrarian Pressure-test the idea. Tell me what could fail, what assumptions are weak, and why this may not work.

2. The First Principles Thinker Strip the idea back to basics. Ask whether I am solving the right problem, or whether there is a better way to think about it.

3. The Expansionist Look for the hidden upside. Tell me how this idea could become bigger, more valuable, or more powerful than I first imagined.

4. The Outsider Look at the idea from a completely different industry, market, or worldview. Show me the blind spots I cannot see because I am too close to it.

5. The Executor Ignore theory. Tell me what I should actually do next, what the first practical step is, and how to test the idea quickly.

Rules:

• Each advisor must respond separately.
• They should not agree just to be polite.
• They should challenge each other where useful.
• Be direct, practical, and honest.

After the five advisors respond, the Chairman must give:

1. Final verdict
2. Biggest risk
3. Biggest opportunity
4. First action I should take
5. Whether I should proceed, improve, pivot, or stop

My idea is:

[PASTE YOUR IDEA HERE]

Why Each Voice Earns Its Seat

The Contrarian exists because confirmation bias is one of the most expensive habits in business. When you are excited about an idea, you naturally filter for reasons it will work. The Contrarian’s job is to find the cracks before reality does.

The First Principles Thinker keeps you from optimising the wrong thing. Many well-executed projects fail because they answered a question nobody needed answered. This advisor drags the conversation back to the actual problem before the team falls in love with a particular solution.

The Expansionist is the counterweight to caution. Once the risks are clear, it is easy to shrink an idea until it becomes safe, small, and forgettable. The Expansionist asks whether the idea could become a repeatable service, a new revenue stream, a stronger client offer, or part of a wider competitive advantage.

The Outsider breaks you out of your own industry’s assumptions. Every field has habits. Some are useful. Many are invisible. The Outsider asks how a software company, logistics firm, startup, sports team, or customer might see the same problem.

The Executor ends the philosophising. Analysis is comfortable. Action is not. This advisor forces the conversation into a concrete next step, a fast test, and a clear measure of whether to continue.

The Chairman does what good leadership must do: synthesise the noise into a decision.

Proceed. Improve. Pivot. Stop.

No fence-sitting.

A Simple Example

Imagine Lumen is considering a new client workshop called “AI Council for Better Business Decisions.”

The Contrarian might say: “Clients may see this as another AI gimmick unless it solves a real business problem. The risk is selling the prompt instead of the decision discipline behind it.”

The First Principles Thinker might say: “The real problem is not that clients lack AI tools. The real problem is weak decision-making: poor assumptions, polite meetings, unclear next steps, and no fast test.”

The Expansionist might say: “This could become more than a workshop. It could become a repeatable consulting framework for CRM strategy, AI automation, sales process design, and executive decision-making.”

The Outsider might say: “Strategy firms use war rooms. Product teams use design sprints. Investors use investment committees. Lumen could position this as an AI-powered decision room for business leaders.”

The Executor might say: “Do not build a full product yet. Test it internally on one Lumen decision, then run a 60-minute pilot with one trusted client. Measure whether it produces clearer risks, better options, and a faster next action.”

The Chairman might conclude: “Proceed, but only as a pilot. The biggest risk is positioning it as an AI trick. The biggest opportunity is turning it into a practical decision-making framework. First action: test it on one real decision this week.”

That is the value of the AI Council. It does not just produce an answer. It improves the thinking around the answer.

This Is a Thinking Tool, Not a Decision-Maker

One important caveat is central to how we see AI at Lumen: the council does not decide. You do.

The point is not to outsource judgement to a model. The point is to arrive at your judgement after genuinely considering the angles you may otherwise have skipped.

AI can surface better questions. It can expose weak assumptions. It can simulate disagreement. It can help you see the upside and downside faster.

But accountability stays with you: to your client, your team, and your own standards.

AI is at its best not when it thinks for you, but when it makes you think harder.

Why This Fits Lumen Business Culture

The AI Council fits Lumen Business Culture because it supports the behaviours strong companies need most: listening, challenge, humility, initiative, learning, and execution.

It gives structure to disagreement without turning it into politics. It helps teams challenge the idea without attacking the person. It makes weak assumptions visible early, when they are still cheap to fix.

That is the right way to use AI in business: not as a shortcut around judgement, but as a discipline that makes judgement stronger. This is why AI belongs inside a strong business culture, not outside it.

Try It

Take a decision you are sitting on right now: a pricing change, a new service line, a hire, a client proposal, a CRM workflow, an AI consulting and system implementation project, or a strategic investment you are unsure about.

Run it through the council.

Read the five voices.

Notice which advisor says the thing you did not want to hear.

That advisor is usually the one worth listening to.

That is what good thinking looks like in the age of AI.

And it is how we work at Lumen.