// DIAGNOSTIC · INTERACTIVE

AI Board Diagnostic. Where your board stands on the nine AI questions.

Ask a CEO where the company is on AI and most boards get an answer in aggregate. This diagnostic takes the nine questions from the AI and the Board guide and scores how far your board has got with each one, from not discussed to governed. You get a one-page read: where the board stands, the question to fix first, and what to do next.

Nine questions · five stages · about ten minutes · nothing stored, no email

01 · The five stages

One scale for every question.

Every question is scored on the same five stages. They separate saying from doing.

0

Not discussed

The board has not asked the question.

1

Discussed

It has come up at the board, but nobody owns it and nothing is written down.

2

Owned

A named executive owns it and the work has started.

3

Measured

The artefact exists and reaches the board on a set cadence, with numbers.

4

Governed

Measured over more than one period, decided on, and could be shown to a buyer next week.

The rule that keeps it honest. A policy or a plan on its own never scores above Discussed. Owned needs a named person. Measured needs the artefact itself, not its title.

02 · Score your board

Answer for one board.

Start with where the company uses AI, then pick the stage your board has reached on each of the nine questions. Pick the stage you could show evidence for today, not the one you are working towards.

// 00Where does the company use AI?

Tick all that apply.

// 01The map

Which workflows use AI today, at what risk, which are agent-ready, augment or human-only, and which one moves this quarter?

Governed looks likeA current operating model with a risk score and an AI class on every workflow, reviewed by the board each quarter, with a named workflow moving this quarter.

Does not apply to this company. Scored out of the questions that do.
// 02New money

Does AI make money the company could not have made before?

Governed looks likeNew AI revenue is reported to the board separately from AI cost savings, every quarter, with at least two periods of history.

Does not apply to this company. Scored out of the questions that do.
// 03Funding new bets

How are new AI bets funded, and who owns them?

Governed looks likeA ring-fenced budget outside the efficiency budget, one named owner, a quarterly review and a written stop point for each bet. At least one bet has been stopped or moved up a stage on that evidence.

Does not apply to this company. Scored out of the questions that do.
// 04The inventory

Can the company produce a list of every AI tool on demand?

Governed looks likeEvery AI tool in every part of the business, with its owner, the data it can see and the date it was last reviewed, produced within a day when asked and checked at least twice a year.

Does not apply to this company. Scored out of the questions that do.
// 05Reconstructing a decision

Could a non-engineer reconstruct any decision the system made in the last twelve months?

Governed looks likeEvery call, input, output, model version and prompt version is logged where the company controls it, and a replay has been run for the board or an auditor.

Applies if the product or the company's own operations use AI to make or shape decisions.

Healthcare note: if the product informs a diagnosis or treatment decision, the evidence bar is higher, and how the software is classified decides it. This tool does not give regulatory advice.

Does not apply to this company. Scored out of the questions that do.
// 06The orchestration layer

Who owns the layer that decides which model or component does what?

Governed looks likeThe company owns the orchestration and the audit trail, can name which tasks run on which components, and has a tested plan for changing model vendor.

Applies only if AI is in the product.

Does not apply to this company. Scored out of the questions that do.
// 07Data rights

Can the company evidence its rights to the data its AI was trained on and uses?

Governed looks likeA data record for each AI feature (sources, licences or consents, retention), kept current and reviewed before each new feature ships.

Applies only if AI is in the product.

Healthcare note: if the product informs a diagnosis or treatment decision, the evidence bar is higher, and how the software is classified decides it. This tool does not give regulatory advice.

Does not apply to this company. Scored out of the questions that do.
// 08AI claims

Is there evidence behind every AI claim the company makes?

Governed looks likeA substantiation file for every AI claim on the website, in proposals and in the pitch deck (autonomous, AI-powered, any accuracy figure), checked before each release and before a raise or sale.

Healthcare note: if the product informs a diagnosis or treatment decision, the evidence bar is higher, and how the software is classified decides it. This tool does not give regulatory advice.

Does not apply to this company. Scored out of the questions that do.
// 09Regulation and standards

Does the board know which AI rules apply to the company, and which framework it runs its AI work to?

Governed looks likeA written view of which AI rules apply and from when (for example the EU AI Act, where the company sells into or operates in the EU), mapped to a recognised framework such as ISO/IEC 42001 or the NIST AI RMF, and reviewed by the board whenever the rules change.

This tool does not give legal advice. It asks whether the board has a written view, not whether that view is right.

Does not apply to this company. Scored out of the questions that do.

0 of 8 answered

// Your board's read

Answer the profile question and every question that applies, and your board's read appears here.

Fix first

What fits

Read the AI and the Board guide Run the deeper version in Claude

Nothing you entered has been stored or sent.

// Take it with you

Whitepaper and Claude prompt

Two artefacts to run this with your chair or your board. The whitepaper sets out the nine questions, the five stages and what governed looks like for each. The Claude prompt runs the diagnostic from your real evidence and ends with a 90-day plan.

// Whitepaper · PDF

Nine questions, five stages

How a board oversees AI without becoming technologists. The nine questions, what owned, measured and governed look like for each, a worked example and how to run it every quarter.

Read the whitepaper→
// Claude prompt · Interactive

The AI Board Diagnostic, run live

Paste the prompt into a Claude project. It takes you through the nine questions one at a time, scores each from your evidence rather than your answer, and ends with the read and a 90-day plan.

Copy the prompt→

This tool is part of the field guide AI and the Board: the questions a board should ask about AI, and the tools that help answer them.

Read the guide

FAQs

What is the AI Board Diagnostic?

A free self-assessment that scores a board on the nine questions it should ask about AI every quarter, from the Ortent guide AI and the Board. Each question is scored on five stages: not discussed, discussed, owned, measured and governed. The result shows where the board stands on each question, the one to fix first, and what to do next.

Who should fill it in?

A chair, NED, PE operating partner or CEO, answering for one board. It works best when two people fill it in separately and compare. Where one answer is lower than the other is where the board conversation should start.

What do the five stages mean?

Not discussed: the board has not asked. Discussed: it has come up, but nobody owns it and nothing is written down. Owned: a named executive owns it and the work has started. Measured: the artefact exists and reaches the board on a set cadence, with numbers. Governed: it has been measured over more than one period, the board has taken a decision on it, and the company could show it to a buyer next week.

Why does a policy not count?

A policy says what should happen, not what does. A board needs to see the artefact: the inventory, the decision log, the substantiation file. A policy or a plan on its own scores Discussed at most. Owned needs a named person, and Measured needs the artefact itself, not its title.

What if some questions do not apply?

The first question asks where the company uses AI. If AI is not in the product, the questions on the orchestration layer and data rights do not apply. If AI is only planned, the question on reconstructing a decision does not apply either. The result is then scored out of the questions that do apply.

Is anything I enter stored?

No. The diagnostic runs entirely in your browser. Nothing is stored or sent, and there is no email field. You can save the result as a PDF from the result panel.

How often should a board run it?

Every quarter, with the same questions in the same order, so the board can see which answers improve. One quarter without a good answer is a process problem. Two quarters is an ownership problem.

// GET IN TOUCH

Clarity when it counts.

If the read shows gaps your board wants closed before the next raise, sale or audit, a 30-minute call is the quickest way to see whether Ortent is the right fit.

Book a 30-minute intro call