How to Present Your AI Strategy to Your Board
To present your AI strategy to the board, bring one page, not a technology tour. Show where the company stands on AI today, the two or three bets you are making, the number each one should move, and the decisions you need from the board. Directors want to judge your choices, not learn how the models work.
Most boards are not yet set up to do that judging on their own. Deloitte's survey of 695 board members and executives found that 66% say their boards have limited to no knowledge or experience with AI, and 31% say AI is not on the board agenda at all. In Corporate Board Member's 2026 What Directors Think survey, only 8% of directors report strong AI expertise on their board, and 40% rank technology developments, including AI, among the hardest issues to oversee. So the quality of your board's AI conversation depends almost entirely on what you bring into the room.
This post covers what directors actually ask about AI, a one-page update you can use at your next meeting, how to build it step by step, and the mistakes that turn an AI update into a demo.
Why does the board conversation about AI matter?
Because boards that talk about AI regularly tend to sit at companies getting more out of it. Protiviti and BoardProspects surveyed 772 board members and executives at the end of 2025. 63% of high-ROI organizations put AI on every board agenda, against 13% of low-ROI organizations. Across the whole sample, only 26% of boards discuss AI at every meeting.
That is a correlation, not proof that agenda time causes returns. But it fits a simple logic. When AI is a standing item with numbers attached, the CEO has to make choices every quarter, and the choices are where the value comes from. When AI shows up once a year as a trends slide, nothing has to change.
For a company of 25 to 500 people, the board is often small: a few investors, a founder, maybe one independent director. That makes this easier, not harder. You do not need a governance framework. You need a clear page and a regular slot on the agenda.
What does a board actually want to know about AI?
Strip away the jargon and directors ask five questions:
- Where do we stand? An honest picture of today, not the ambition.
- Where are we betting, and why there? Which problems, and which numbers they should move.
- What will it cost, and when will we see it? Money, people's time and time to results.
- What could go wrong? Data, customers, quality of output, and your people.
- What do you need from us? An approval, a budget, an introduction or a decision to stop something.
Gartner's guidance for executives presenting AI to boards makes a similar case: lead with shareholder value, risk and competitive position, be candid about maturity and skills gaps, and avoid overpromising. Notice that none of the five questions is about which model or vendor you use. If your deck spends more time on tools than on these five, rebalance it.
What should a one-page board AI update include?
A board AI update is a single page that shows your AI position, your bets and the decisions you need, in the same format every quarter. Keeping the format fixed is the point: it lets the board see movement instead of a new story each time. Here are the six parts:
- Your stage. One line on where the company sits in the four stages of AI maturity: Scattered, Experimenting, Operating or Compounding. Our AI maturity model explains each one. Add what it would take to reach the next stage.
- The portfolio. No more than three bets. For each: the problem, the owner, the metric, the baseline, the latest result and a status of scale, fix or stop.
- What, How, Build. One line per bet on the approach: what work is changing, how you will do it, and whether you will configure tools you already own, buy a product or build something custom, with the reason.
- Spend and capacity. Money committed, and the hours per week it takes from named people. In a mid-sized company, capacity is usually the real limit, and boards rarely see it.
- Risks and guardrails. The two or three that matter: what data can go into which tools, where a person reviews AI output, and whether you have a written AI policy.
- Decisions requested. One to three specific asks, each with a yes or no answer.
The difference between a weak update and a strong one is easy to spot:
- Weak: a list of tools, a dozen pilots, market statistics and "promising early results."
- Strong: one stage, three bets with owners and numbers, a cost in money and hours, and a clear ask.
How do you build the update, step by step?
Plan on two to three hours the first time, most of it gathering numbers. After that, a quarterly refresh takes far less.
- Start from the decision you need. Write the asks first. If there is nothing for the board to decide, it is a status report and it can be an email.
- Place yourself honestly on the four stages. Ask your leadership team three questions: who owns AI, where does it show up in real work, and what number has it moved? If the answers spread across two stages, pick the lower one.
- Cut the list to three bets. Directors cannot weigh twelve pilots, and neither can you. Everything else goes on a single line, such as "four small experiments, under budget, no decision needed."
- Fill in the numbers row by row. Baseline first. If a bet has no baseline, say so, and make capturing one part of the plan. Our guide on how to measure AI ROI sets out the method.
- Name the capacity cost. For each bet, write who spends how many hours a week on it. This is the line boards most often miss, and the one most likely to explain a delay later.
- Write the risk lines in plain words. "Customer data does not go into public chatbots, and the team has an approved tool for that" beats a paragraph about responsible AI.
- Test it with one director. Send the page a few days early to the director most likely to push back, and ask what is missing.
- Close with a dated calendar. List what happens before the next meeting, with dates and owners. Next quarter, you report against it.
How do you show progress quarter after quarter?
Keep the page and change the numbers. The board should be able to set this quarter's page next to last quarter's and see three things at a glance: did the stage move, did any bet change status, and did the numbers move?
Make stop decisions visible. A bet you stopped, with a clear reason, is a sign of discipline, not failure. A board learns to trust a CEO who stops things faster than one who reports that everything is on track.
The right measure also changes as you move up the stages:
- Experimenting: report what you learned and what you stopped.
- Operating: report results against KPIs the board already tracks, such as quote turnaround or days to close the books.
- Compounding: report unit economics, such as cost per order or revenue per employee, quarter over quarter.
What mistakes should you avoid?
- The demo. A live demo is fun and tells the board little about value. If you show one, keep it to two minutes and tie it to a bet on the page.
- The trends slide. Market statistics about AI are not your strategy.
- The pilot list. Twelve pilots read as twelve decisions you have not made.
- Hiding the capacity cost. If your best operator spends a day a week on AI, the board should know, because it explains what else slows down.
- Handing it to IT. If IT presents AI, the board will treat it as an IT topic. IT can explain how something is done safely. The strategy is yours to present.
- Overpromising. In the Corporate Board Member survey, 38% of directors named deploying AI across the business as a top priority for the year. That pressure makes it tempting to promise big numbers early. Promise a learning plan with dates instead, then beat it.
Where should you start before your next board meeting?
Draft the page this week, with blank cells where you do not have numbers yet. The blanks are your work plan. Share it with your leadership team, fill in what you can, and send it to the board with the pre-read. One page, three bets, clear asks.
If you would like an outside view of where your company stands before you take it to your board, Journey offers a free CEO AI audit to help you see where you are and what to do next.
Frequently Asked Questions
Bring one page instead of a long deck. Show your current stage of AI maturity, no more than three bets with an owner, metric and baseline for each, the money and hours they take, the main risks and guardrails, and one to three decisions you need. Keep the same format every quarter so directors can see what moved.
For most companies, at every regular board meeting, even if only for ten minutes. A Protiviti and BoardProspects survey found 63% of high-ROI organizations discuss AI at every board meeting, compared with 13% of low-ROI ones. That does not prove cause and effect, but a standing slot forces regular decisions about what to scale and what to stop.
Report the business number each AI bet is meant to move, such as quote turnaround, days to close the books or cost per ticket, with its baseline and latest result. Add money spent and hours of staff time used. Early on, also report what you stopped and why. Avoid usage counts or logins on their own, because they say nothing about value.
Not necessarily. Few boards have one: only 8% of public company directors report strong AI expertise on their board. What a smaller board needs more is a CEO who brings a clear, honest update every quarter. If AI is central to your strategy, an advisor or a director with hands-on AI experience can help, but it does not replace management ownership.
The CEO should present the AI strategy, because the board needs to judge business bets, trade-offs and priorities, which are CEO decisions. IT or a technology lead can join to cover security, data access and how tools are run safely. If IT presents the whole topic, directors tend to treat AI as a technology project instead of a strategy question.
