Build a personal AI boardroom for the decisions you make alone

The hardest part of running a business alone is not the workload. It is making consequential decisions with no one in the room to push back. Should you take the awkward client? Drop the underperforming service? Sign the lease? A bigger company would put that question in front of a few people whose job is to see what you cannot, the finance head who flags the cash risk, the lawyer who spots the clause, the colleague who asks the obvious question you were too close to ask. A solo owner makes the same call at the kitchen table, and the only voice in the room is the one that already agrees with them.

AI gives a one-person business a cheap version of that room. Allie K. Miller, one of the most widely followed voices on practical AI in business, coined the idea of a personal AI boardroom: you spin up several AI personas, a sceptical finance head, a cautious risk voice, and make them stress-test your thinking before you commit. It is not a decision-maker and not an oracle, but a panel of advisers you can convene on demand. Miller is equally blunt that the tool underneath is exactly that, a tool: "a tool (not a silver bullet)," in her words. Used as a sounding board rather than an answer machine, it earns its place.

The lonely-decision problem

Decisions made alone share a predictable weakness. You bring one perspective, usually the one that wants the decision to be easy, and you bring all your blind spots with it. There is no friction. Nobody asks the cost question, the what-if-this-goes-wrong question, the how-would-the-customer-feel question. The decision feels considered because you thought about it hard. Thinking hard from a single angle is not the same as being challenged from several.

The fix a board provides is not intelligence. It is distinct viewpoints, each with its own thing to worry about. You can reproduce that with AI by giving the model a role to play and the context to play it well.

What a persona board is

A persona board is a short, named set of advisers you ask the model to become, one at a time, when you are weighing a real decision. You describe the decision, hand over the relevant facts, and then ask each adviser to react in character: what they would worry about, what they would want to know, where they think you are fooling yourself.

The trick is to keep the roles few and sharply different, so each one pulls in a genuinely separate direction. Three is usually enough. Greg Shove, who argues that leaders should consult AI before big calls and treat it as a kind of standing adviser, makes the case that the habit only works if you actually use it, "model the behaviour" yourself rather than admiring it from a distance. A board you convene twice a year is theatre. A board you bring out for every real decision becomes a discipline.

Three advisers worth keeping

You can invent advisers to suit your business, but three earn their place in almost any firm.

The sceptical finance head. Brief this one on the numbers and ask the uncomfortable money questions. What does this really cost once everything is counted, beyond the headline number? What has to be true for it to pay off? What happens to cash if it takes twice as long as you hope? A solo owner in love with an idea needs this voice most, and listens to it least.

The cautious risk voice. This adviser reads for what could go wrong: the obligation you are taking on, the client who might not pay, the promise you cannot keep, the way this looks if it ends badly. It is not there to talk you out of things. It is there to make sure you walk in with your eyes open.

The customer. The voice most often missing from a solo decision is the person you serve. Ask the model to react as a specific, well-described customer: how this change lands for them, what they would quietly resent, what they would happily pay more for. It keeps the decision pointed outward, at the people who actually fund the business.

A good board does not tell you what to do. It tells you what you missed.

Briefing them with your context

A board is only as good as what it knows. Ask the model to react cold and you get generic advice that fits any business and helps none. The difference is context. Before you ask for reactions, give each adviser the specifics: your size, your margins, the actual numbers, the real client, the constraint that makes your situation yours. Miller's own short list of how to get value from these tools starts exactly here: "know your business," then experiment, then refine what you ask for. The same prep that makes you a better decision-maker makes the board a better adviser.

A practical habit: keep a short written description of your business, the numbers and constraints that rarely change, and paste it in at the start of every board session. The advisers stop guessing and start reacting to your actual situation.

There is a second move that sharpens the board further: tell each adviser to argue, not to soothe. Left to its own devices the model wants to be agreeable, and an agreeable board is worthless. So instruct the finance head to find the reason this fails, the risk voice to name the worst plausible outcome, and the customer to say the thing they would never say to your face. You are not looking for permission. You are looking for the objection a friend would be too polite to raise, and the model, briefed properly, will raise it without flinching.

The honest limits

A persona board is a mirror, not an oracle. It reflects your framing back at you with more discipline than you would manage alone, which is genuinely useful, but it does not know things you have not told it and it cannot want the right outcome for your business the way a real adviser can. If you brief it to confirm what you already believe, it will oblige, fluently. The value comes from asking it to attack the decision, not defend it.

It is also no substitute for a real human when one is available and the stakes are high. For the biggest calls, the lease, the hire, the bet-the-business decision, the board is a way to arrive better prepared at a conversation with an actual accountant, lawyer or mentor. Use it to sharpen the questions, not to skip the people.

What to do this week

Take one real decision you are sitting on. Write down the facts in a few lines. Then convene the three advisers, one at a time, and ask each to tell you what worries them and what they would want to know before saying yes. Read what comes back not for the answer, because there isn't one, but for the question you had not asked yourself. Most owners find at least one, and that is the whole point.

Making lonely calls with more rigour is a large part of what the AI Lessons for Leaders sessions build: practical ways for a founder to use these tools as a genuine thinking partner rather than an answer machine. If that is the habit you want, book a personal lesson and we will set yours up around the decisions you actually face.

Sources and further reading

• Allie K. Miller's "personal AI boardroom," in Coalition's write-up of her Activate 2026 talk, 16 April 2026. Where the idea comes from: spinning up AI personas to stress-test a decision before you commit to it.
• Allie K. Miller, a LinkedIn post of plain AI advice for business owners, November 2023. Her line that AI is "a tool (not a silver bullet)", and the know-your-business-then-experiment sequence.
• Greg Shove, gregshove.com, 2025. His case that leaders should treat AI as a standing adviser, and actually use it themselves.