Your board, investors and bank want four things from you on AI: what you use it for and who is accountable, what it has changed, what has gone wrong, and what you need them to decide. Put those four answers on one page every quarter. I call it the Quarter Page, and it is below, ready to copy.
Few boards ask for it yet, so an owner who brings it first sets the standard for the room. This applies whether you have a formal board, a two-person advisory group, an outside investor, a co-owner or a bank manager who reviews your numbers. If anyone has the right to ask "what is going on with AI in this business?", this is how you answer before they ask.
What do boards and investors actually ask about AI?
They ask less than you would expect, and they measure even less. Most of the evidence on board oversight comes from large companies, so I have set it beside the two other audiences a small business answers to, checked at source on 3 November 2026.
| Who is asking | What the evidence shows | What it means for you |
|---|---|---|
| Boards of large listed US companies | In March 2026, ISS STOXX reviewed 3,048 US companies in the Russell 3000 and S&P 500 (data from January 2026). Only 245 (8%) disclosed board-level oversight of AI, and 275 (9%) said they had established AI policies. | Disclosure is the exception even at companies with a governance team. Being informal is normal. Being unable to answer is the risk. |
| Boards of private companies | The 2025 NACD Private Company Board Practices and Oversight Survey (117 people who serve on or support private company boards, fielded May 8 to June 2, country not stated). Of the 65 who answered the AI activities question, 40 (about 62%) had set aside agenda time. Only 11 (about 17%) had audited where AI is used, and 10 (about 15%) had set metrics for management reporting. | Boards talk about AI far more than they measure it. A page of facts stands out. |
| Banks and lenders | In September 2026, the San Francisco Fed's Economic Letter 2026-25 compared 1,006 US banks. Banks with high AI intensity made a much smaller share of their loans to small and medium-sized businesses: 12% against 21%. | The authors say these are correlations, not proof of cause. Still, a lender weighing hard data rewards a business that can show its records. |
The verdict: talking about AI is common, evidence about AI is rare. That gap is your opportunity, because a short page of facts is more than most of the people asking for it have seen.
Why bring a board page if you have no board?
Because the reader is whoever holds you to account, and in a small business that is rarely a formal board. It might be a co-owner, a trusted adviser, an investor with a small stake, the person at the bank who reviews your accounts, or your own senior team.
Law firm Alston & Bird, writing in Law360 in June 2026 about public company boards, put the starting point plainly: "Effective oversight begins with visibility, which requires a comprehensive catalog of what AI tools and systems the organization is using, whether developed internally or provided by third parties." That sentence works at any size. If you cannot list what is in use, nobody can oversee it, including you.
In the UK, the Institute of Directors makes the same case from the other direction in its business paper on AI governance in the boardroom: boards that move beyond compliance toward foresight and stewardship are best placed to get value while protecting trust and reputation. You do not need a board to borrow the habit.
If you have no one who holds you to account, invite two people for one hour a quarter and give them the page. That is a board for a business your size.
What goes on the Quarter Page?
The Quarter Page is four questions on one sheet, answered the same way every quarter. Here is the locked wording, then the template.
The Quarter Page is a single page an owner brings to every board, advisory or investor meeting once a quarter, with four headings: What are we using AI for now, and who is accountable for each use? What has it changed, in numbers or in plain facts? What went wrong or nearly did, and what did we do about it? What are we asking you to decide? A business that can fill the page truthfully four quarters in a row is being governed. One that cannot is only experimenting.
Copy this table into a document and fill the right-hand column.
| Heading | The question | What to write | Evidence to attach |
|---|---|---|---|
| 1. Used | What are we using AI for now, and who is accountable for each use? | One line per use: the job, the tool family, the named owner, and what data goes in. | The list your team keeps of tools in use (see shadow AI). |
| 2. Changed | What has it changed, in numbers or in plain facts? | One or two measures, a baseline and today's figure. If you have none, say so and name the first one to start. | The before and after, however rough (see how to measure whether AI is working). |
| 3. Slipped | What went wrong or nearly did, and what did we do about it? | Each slip in one sentence, who caught it, and the check you added. "Nothing" is only acceptable if someone is actually looking. | Your log of checks and corrections (see training staff to check AI output). |
| 4. Next | What are we asking you to decide? | One decision, with a date: start, stop, widen, or pause. | The policy, budget line or supplier card the decision touches. |
Notice what the page leaves out: tool comparisons, vendor claims and a list of everything AI could do. The reader wants to know what is happening in your business, who owns it, and what you need. This is my governing principle at work: outsource the doing, not the thinking. AI can draft the first version of your answers, but the thinking behind them has to be yours.
What does a filled-in Quarter Page look like?
Take a hypothetical 40-person building supplies distributor with a three-person advisory board. The figures below are invented to show the shape, not to suggest results.
| Heading | The owner's entry |
|---|---|
| 1. Used | Drafting first replies to customer emails (customer service lead, no customer account numbers pasted in). Drafting quotes from a standard template (sales office manager). Trial: matching supplier invoices to delivery notes (finance manager, no bank details). |
| 2. Changed | Quote turnaround averaged three working days in the spring and is now one, measured from the quote log. Invoice matching trial: not yet measured, first measure agreed for next quarter. |
| 3. Slipped | One drafted customer reply quoted a discount that had ended. The customer service lead caught it before sending. Check added: every reply that mentions a price is read against the current price list. |
| 4. Next | Decision requested: approve moving invoice matching from trial to routine from January, if the first quarter's mismatch rate is below the manual rate. |
That page took the owner perhaps an hour. It tells the advisory board where AI is used and who answers for it, gives them one measure to trust, shows that mistakes get found, and ends in a decision they can actually make. If a bank manager or an investor asked for a briefing on AI tomorrow, this page is the briefing.
The Mistake I See Most Often
Owners bring a success story with no slip in it, or a list of tools with no decision at the end. In the business teams I train, the pattern is consistent: people describe what the tools can do, and the questions that matter are about ownership and evidence. Who is responsible for this use? How would you know if it were wrong?
The NACD survey points the same way. Asked which single factor is most critical to effective AI oversight at the board level (84 respondents), the top answer was director education and expertise (about 27%), followed by a management team with the right AI focus (25%), then enough agenda time (about 20%). Appropriate information and metrics came fourth, at about 17%. The page fixes that last one for you, and it does the first one quietly, because anyone who reads four of them has learned how AI is used in your business.
A page with a slip on it is stronger than a page without one. A reader who sees a caught mistake and an added check trusts the other three headings more.
What will your bank or lender look for?
Expect them to weigh records, not stories. The San Francisco Fed authors explain the likely mechanism in their summary: "AI helps banks in processing hard data, such as credit scores and financial statements, and issuing fewer small business loans that rely more on soft information, such as personal and business relationships." They also write that "GenAI might eventually assist banks in processing soft data, which could bridge the information gap for SME lending." Their AI measure is built from job postings, so it reflects mostly in-house AI work, and the study cannot say that AI caused the pattern.
So treat this as a reason to keep dated, written records, not as a prediction about your own bank. A business that can show what it uses, what it measured and what it fixed is easier to lend to on hard information, and just as easy to back on relationship. The page is the same document either way. It also pairs with the questions you ask your own suppliers: see what to ask suppliers who use AI and telling customers when you use AI.
If you are planning a sale, a longer document serves a different purpose: the Sale-Ready File is for handing a business over, while the Quarter Page is for running it.
What should you do this week?
You can do this in about an hour with no data team.
- List every use of AI in the business today, with a named owner for each. If you are not sure, ask the team, and read is your team really using AI first, because the real list is usually longer than the official one.
- Pick one measure for your most important use. Write down today's figure, even if it is a rough estimate, and label it as one.
- Write down one slip, however small, and the check you now run. If you cannot think of one, ask the person who uses the tool most.
- Write one decision you want, with a date.
- Put the four entries into the Quarter Page table and send it to your co-owner, adviser or bank contact before the next conversation.
- Put the page in the calendar for the same week each quarter, and check it against your AI policy and your AI plan.
If you want a test before you start, I use the Readiness Test: three questions an organization must be able to answer consistently, from the top to the front line, before it can call itself AI ready: What is AI for here? What may I do with it? Who decides when it goes wrong? Your page should let a reader see all three answered.
Where to go next
If you want the page to be true across the whole team, not just on paper, my AI training for teams is built for exactly that. To see where your business stands before you write it, take the Business AI Readiness Scorecard. And if you are planning a leadership event or an annual conference, my business keynotes are built around questions like these.
This article is general information, not legal, financial or investment advice. For decisions on director duties or lending, speak to a qualified adviser.
Sources and further reading
- Hinza Zeru and Andrew Schultz, ISS STOXX, Mind the Governance Gap: The State of Board Oversight and AI Policy in U.S. Companies, 3 March 2026. Opened 3 November 2026.
- National Association of Corporate Directors, 2025 NACD Private Company Board Practices and Oversight Survey, AI oversight activities conducted by the board and critical factors for effective AI oversight at the board level, data pack dated 26 August 2025. Opened 3 November 2026.
- Greeshma Avaradi, Naomi Halbersleben, Zheng Liu and Mark Spiegel, Federal Reserve Bank of San Francisco, How AI Adoption Might Affect Bank Lending, FRBSF Economic Letter 2026-25, 21 September 2026. Opened 3 November 2026.
- Courtney Quirós, Cynthia J. Cole and Lex Diktas Mayo, Alston & Bird, How Boards Can Shrink the AI Governance Gap, originally in Law360, 16 June 2026. Opened 3 November 2026.
- Institute of Directors, AI Governance in the Boardroom, IoD Business Paper (no publication date shown on the page). Opened 3 November 2026.
Dan Fitzpatrick, The AI Educator, is a keynote speaker, Forbes contributor and five-times bestselling author who has trained more than 150,000 educators in 30+ countries. About Dan.
