Board governance

How boards should evaluate an AI strategy

A board does not need to understand the technology. It needs to understand what the strategy is actually for, and five tests separate a real plan from a slide deck.

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

You do not need to understand the technology to judge an AI strategy. You need to understand what the strategy is for. Most board packs in front of me are about what the technology can do. The question a director should be asking is what the business is going to do with it, who answers for it, and what it costs if it fails.

This guide gives you the five tests to run on any AI strategy put in front of your board, what the research says about boards today, and where the process usually goes wrong.

What an AI strategy actually is

An AI strategy is not a list of tools. It is a decision about where the business will let machines make decisions, and where it will not. If a document cannot tell you that in a few sentences, it is not a strategy yet.

I have sat through more than one board where the strategy was a shopping list. We will use AI in customer service. We will use AI in marketing. We will automate reporting. That is not a strategy. It is an ambition without an owner, a cost without a budget, and a risk without a name.

Two thirds of those transformations did not deliver. The strategy was not the problem. The discipline around it was.

The five tests every strategy should pass

Fuzzelogic works with a definition of AI-ready that comes down to five things being true. Run them against any AI strategy and you will see within an hour whether it is real.

1. Reachable. Can the data the AI needs actually be found when it needs it?
2. Trustworthy. Do you know the data is accurate, current, and complete?
3. Explainable. Can someone explain why the system made a particular decision?
4. Changeable. Can the system be changed when the business changes?
5. Governed. Has someone decided what the system may and may not do?

Take each test and ask the room the direct question. Where is the data? Who last checked it? Who can explain this decision to a customer? If the strategy gets vague at the same point every time, that is your real problem, not the technology.

What the research says about boards

The gap is not in the technology. It is in the boardroom. Two findings stand out.

A board that is rushed and uninformed in equal measure tends to do one of two things. It either waves everything through on a vendor slide, or it blocks everything out of fear. Both are expensive. The first costs money on projects that go nowhere. The second costs the business every year it stands still.

The same research found that 47% of CEOs say they personally lead AI implementation. That is worth pausing on. If the chief executive personally leads it, who challenges it? A strategy needs at least one director whose job is to ask the awkward questions.

What to look for when the strategy is presented

When the presentation is over, the real test begins. Ask for three things.

First, the owner. Name the person who answers when it goes wrong, not a committee. Committees do not lose sleep. People do.

Second, the money. The strategy should state what it costs, what it saves, and what it earns, in numbers a finance director would accept. If the benefits are given as percentages without a starting point, ask for the starting point.

Third, the off switch. What is the criteria for stopping, and who has the authority to pull it? The best strategies are the ones that admit they might be wrong.

Where boards go wrong

The pattern I see most often is the pilot that never ends. A team runs a small experiment, it shows promise, and everyone declares success while no one scales it. Months pass, costs mount, and the same pilot appears in next year's pack with new branding.

The antidote is the four steps Fuzzelogic uses for responsible adoption: find it, classify it, govern it, train for it. Find where AI already exists in your business, often under the radar. Classify each use by what happens if it fails. Govern the ones that matter. Train the people who have to live with the result.

If a strategy cannot name its governance, it is not ready for a board vote.

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 strategy worth approving is the one that survives your questions. Use the five tests. Name the owner. Count the money. Find the off switch. If a document survives all of that, it probably deserves your vote. If it does not, better to find out now than a year and a budget later.

For the fuller picture, read the one-page AI plan next, 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.