Board governance

The one-page AI plan for boards

Most AI strategies are too long, too vague, and too hard to act on. A one-page plan forces clarity and gives the board something it can actually vote on.

By Zakir Hoosen, Director, Fuzzelogic Solutions. Board-level guidance in plain English.

Boards do not need another 40-page slide deck about AI. They need one page that answers six questions. If a strategy cannot fit on a single page, it is not clear enough to act on. That is not a constraint. It is a discipline.

Board packs grow thicker every year while the decisions inside them grow thinner. The one-page plan exists because brevity forces clarity. If you cannot explain where AI fits in your business on one sheet of paper, you have not thought it through yet.

Why one page works

A long document lets people hide behind the words. You can fill twenty pages with diagrams and still not answer the question that matters: what is this going to do, who owns it, and what does it cost if it fails? One page does not give you that room. Every line earns its place or it does not belong.

The one-page plan is not a summary of a longer document. It is the document. It is the thing the board votes on, the thing the executive team works from, and the thing you revisit every quarter.

What goes on the page

Six sections. No more. Each one answers a single question, and if any section is empty, that tells you something too.

1. The problem. What specific business problem does this solve? Not "efficiency" in general. A named problem with a named owner.
2. The data. What data does the solution need, where does it live today, and who is responsible for making it ready?
3. The approach. What will AI actually do, and what will humans still do? Draw the line clearly.
4. The cost. What does this cost to build, what does it cost to run, and what is the expected return?
5. The risk. What happens if it fails, who carries the consequence, and what is the off switch?
6. The governance. Who oversees this, how often do they review it, and what triggers a stop?

That is the plan. Six lines of six sections. A director should be able to read it in five minutes and ask a meaningful question about any part of it.

The numbers behind the discipline

The reason this discipline matters is that most AI efforts do not deliver what they promise. The research is blunt.

Two thirds of transformations fail to deliver. Boards are rushing and underinformed. The one-page plan exists to slow the rush just enough to ask the right questions.

Who fills it in

This is not a job for the IT team alone. It is not a job for the vendor either. The one-page plan should be completed jointly by the business owner of the problem, the person who understands the data, and the person who will carry the cost if it fails. If those are three different people, good.

The board's job is not to fill in the plan. Its job is to test it. Read each section and ask the direct question. If the answer is vague, send it back.

How it connects to AI-ready

Fuzzelogic works with a five-part definition of AI-ready: Reachable, Trustworthy, Explainable, Changeable, Governed. The one-page plan maps directly to those tests. If the data section cannot show that the data is reachable, it fails the first test. If the governance section is empty, it fails the fifth.

The assessment itself takes two to four weeks, is fixed price, and you own the verdict. The one-page plan is what you build from that verdict.

The risk section is the most important

Boards often spend the most time on the cost section. They should spend the most time on the risk. Cost is recoverable. Reputation is not.

The Palantir white paper on agentic governance makes the point well: an agent that fails 1% of the time might be fine for sending sales emails. It is not fine for code that touches customer money. Classify every AI use by what happens when it fails, and by whether you can reverse the damage.

The Anthropic CISO put it plainly: "Zero risk is not the job." The most effective controls are human approval on consequential actions. Your risk section should name which actions require a human.

The governance section

A plan without governance is an ambition without a referee. The governance section should name the person or committee that oversees the AI use, how often they review it, and what specific trigger causes a stop. "We will monitor it" is not governance. "The CFO reviews the cost and outcome every quarter, and any overage of 15% or more triggers a halt" is.

The honest version

Fuzzelogic is an Isle of Man firm that has spent nineteen years modernising banking, insurance, healthcare, retail, manufacturing, and government platforms. We tell boards what most consultants will not: the honest answer is often that AI should not touch a process at all, and when that is the case, we put it in writing rather than build it anyway.

The one-page plan is not a form to fill in. It is a discipline. If you cannot fill in a section, that section is telling you something. Listen to it. A blank space on a one-page plan is worth more than a full page of vague promises.

The plan worth approving is the one that survives your questions. Six sections, six questions, one page. If it does not fit, it is not clear enough. If it is not clear enough, it is not ready for your vote.

Start with the assessment. Two to four weeks, fixed price, and you own the verdict and the roadmap whether or not we build any of it. When you are ready to talk AI, call Fuzzelogic Solutions and ask for Zak. www.FuzzelogicSolutions.com | info@FuzzelogicSolutions.com | +44 (0)1624 618950

For the fuller picture, read how boards should evaluate an AI strategy first, then who is accountable when AI goes wrong. The full library is on our index. Our site explains how Fuzzelogic approaches AI for business. You can reach Zak directly via our contact page.