To present an AI strategy to a board, frame it as a business decision rather than a technology briefing: state the problem, the options, the expected value with its uncertainty, the risks and how they are controlled, and a specific, limited request with dates when the board will review progress. Boards rarely reject AI because it is unfamiliar; they push back when the ask is open-ended, the costs are incomplete or the risks are glossed over.
Frame the business decision
Open with the decision you need, in one sentence: "We are asking the board to approve £150,000 and six months for two AI pilots in customer service and contract review, with a decision on scaling in March."
Then connect it to the board's own priorities:
- which strategic objective it serves
- what happens if you do nothing, including competitive pressure
- why now
Keep the technology to what the board needs to judge risk and value.
Explain value and uncertainty
Boards trust ranges more than single numbers. For each use case, show:
- the baseline today
- the expected improvement as a range, with the assumption behind it
- all costs, including staff time and ongoing running costs
- what would make it fail, and how you will know early
Our guide to measuring AI ROI has a worked example of benefits, costs and uncertainty.
Cover risk directly: data protection, accuracy, bias, security, reliance on a vendor and the effect on staff. Say what controls are in place and who owns them.
Ask for a bounded commitment with review gates
Avoid asking for a large, multi-year programme up front. Ask instead for:
- a limited budget and timeframe for the first stage
- clear success and stop criteria
- a date when you will return with results and a recommendation to scale, adjust or stop
This makes approval easier and builds the board's confidence for the next stage.
A board-paper outline
- Decision requested (one paragraph)
- Strategic context: the objective, and the cost of inaction
- Proposed use cases: problem, approach, owner
- Value: baseline, expected range, assumptions
- Costs: one-off and ongoing
- Risks and controls: including governance and human oversight
- People impact: roles, training, communication
- Review gates: dates, success criteria, stop criteria
- Appendix: detail for those who want it
Aim for four to six pages plus appendix, and a presentation of no more than ten minutes before questions.
Board questions about AI, and how to answer them
| Likely objection | Helpful answer |
|---|---|
| "What if it gets things wrong?" | Explain human review, accuracy checks and how errors are caught before they reach customers |
| "Is our data safe?" | Set out which data is used, where it is processed, and the contractual and technical safeguards |
| "Why not wait until the technology settles?" | Show the cost of waiting, and that a bounded pilot limits exposure while building capability |
| "What does this mean for jobs?" | Be honest about how roles will change, and describe training and redeployment plans |
| "How do we know the numbers are real?" | Point to the baseline, the pilot design and the review gate where results will be tested |
| "What is the total cost if we scale?" | Give an indicative range, and make clear the board decides at the next gate |
Related reading
- AI strategy subject guide: building, documenting, measuring and presenting an AI strategy
- AI use-case and business-case pack: a worked first-year business case with a downside case
- How to get buy-in for your AI strategy proposal
- How to convince leadership to invest in AI
- AI strategy document template: what to include
Briefing a board and challenging a vendor are core outcomes of our AI strategy course for executives, a non-technical programme covering use cases, ROI, governance and stakeholder alignment.
The bottom line
Present AI as a business decision with a clear ask, honest ranges, real risks and dated review gates. A bounded first step that the board can check is easier to approve than a vision it has to take on trust.