How to make AI a board-level priority without overpromising
Most boards are not ignoring AI. That is not the problem. The problem is that they have seen three flavours of it already: the breathless pitch from a vendor, the POC that never scaled and the strategy deck that sat on a shelf. When the next senior leader walks in and asks for budget and attention, the room has already been burned.
The mistake most organisations at this scale make is conflating advocacy with credibility. They build the case for AI the way a vendor would - emphasising potential, referencing headline figures, showing what is theoretically possible. Boards at £500m to £1.5bn revenue are not short of ambition. They are short of confidence that this time will be different.
This article gives you a practical framework for presenting AI as a board-level strategic priority: how to earn credibility before you ask for commitment, how to frame value in language that resonates with a CFO and a COO simultaneously and how to set expectations that survive contact with reality.
The board is not sceptical about AI - it is sceptical about you
This is an uncomfortable place to start, but it is the right one.
By the time a senior leader is making the formal case for AI investment, the board has usually already discussed it informally. They have read articles. Someone on the NED panel has seen a competitor do something interesting. They are not waiting to be educated on what AI is. They are waiting to see whether the person in front of them understands the business well enough to make it work here.
The credibility gap closes the same way in every engagement we have been involved in: not through better slides, but through sharper problem definition. The leaders who get board buy-in are the ones who arrive with a specific operational problem, a clear articulation of what it costs today and a grounded view of what a different approach would produce.
Consider a COO at a specialist logistics business with £700m in revenue. She did not present an "AI strategy". She presented a problem: the business was losing margin on its last-mile contracts because manual route optimisation was running 48 hours behind real-time conditions. She quantified it - not perfectly, but credibly. She showed what a decisioning system that processed live data would change. The board approved the project in one meeting. That is not unusual when the problem is real and the person presenting it owns it.
The ask: before you write a single slide, write one paragraph that describes the problem in operational terms. If you cannot do that, you are not ready for the board conversation.
Frame value in terms of margin, not capability
AI initiatives fail at the board level for two reasons. The first is vagueness. The second - less discussed - is the wrong type of specificity.
Describing a capability ("the system will process unstructured data and generate automated summaries") tells a board nothing about why they should care. Describing a feature roadmap is a technology conversation. The board needs a commercial one.
The three lenses that work consistently at this level are:
- Margin protection - where is the business leaking money due to manual processes, slow decisions or poor data quality?
- Revenue at risk - what customer behaviours, competitive moves or market shifts are the business currently blind to?
- Operating leverage - where is headcount or cost growing in line with revenue when it should not be?
Each of these connects AI investment to the financial metrics a CFO actually manages. A business intelligence platform that reduces the time from data to commercial decision is not an "analytics investment" - it is a tool that gives sales and category teams the information they need before margin decisions close. That framing changes the conversation.
A PE-backed retail business in the £800m range was preparing for exit. Their data function was producing accurate reports, but the cadence was weekly and the outputs were static. The commercial director could not answer questions from the acquirer's DD team without a three-day turnaround. The problem was not data quality - it was decision latency. Framing the investment in those terms got it approved as part of the exit readiness programme. The same investment, described as "modernising the BI stack", had sat unapproved for eighteen months.
Set governance expectations before you set ambition targets
The single fastest way to lose board confidence after winning it is to return six months later with scope creep, missed timelines and a request for more budget. This is so common it has become a cliché. It is also entirely avoidable.
Boards at this size are not opposed to iterative delivery - they are opposed to ambiguity dressed up as agility. There is a meaningful difference. Before you ask for a mandate, the board needs to understand four things:
- What does success look like at 90 days, not just at project end?
- Who owns the outcome - not the project, the outcome?
- What would cause you to stop or change direction, and who makes that call?
- How will we know if this is not working before we have spent the full budget?
This is not a framework for managing failure. It is a framework for demonstrating that the person asking for investment has thought seriously about accountability. That posture - more than any ROI model - is what converts sceptical board members.
If you are working with an external partner, the same questions apply. A credible consultancy should be able to answer all four before any contract is signed. If they cannot, that is information worth having.
Build the roadmap in phases, not horizons
"Horizon one, two and three" planning has its place. It does not have a place in an early-stage AI investment conversation with a board that has not yet committed.
Phased delivery, by contrast, works because it maps investment to evidence. The structure that tends to hold up in practice looks like this:
Phase one: diagnostic and problem validation (four to six weeks, low cost). Confirm that the problem is real at the scale you claimed. Quantify the baseline. Identify the data assets and gaps. This phase costs a fraction of the total investment and removes the largest source of downstream risk: starting with the wrong problem.
Phase two: contained build and measurement (eight to twelve weeks). Solve one version of one problem. Not a pilot for the sake of optics - a production-ready solution to a scoped use case. Establish the measurement framework at the start, not the end.
Phase three: scale and operationalise. Once phase two has produced measurable results, the conversation about scale is no longer a pitch - it is a proposal grounded in evidence the board already trusts.
This structure also makes the financial exposure manageable. A CFO who is unconvinced about AI will approve a £50k diagnostic far more readily than a £500k programme. And if the diagnostic reveals that the original problem was misframed - which happens - you have saved the business significant money and earned credibility in the process.
What happens if you get the board case right
Getting AI to the board agenda is not the goal. Getting it resourced, governed and delivered in a way that compounds over time - that is the goal.
The leaders who do this well do not wait until they have a perfect business case. They start with a real problem, they frame it commercially, they set honest expectations and they build a track record that makes the next conversation easier than the last.
The cost of getting this wrong is not just a failed project. It is the organisational debt that accumulates when the board associates AI with wasted spend. That debt is hard to retire. We have seen businesses where a single poorly managed initiative set the data agenda back by two or three years - not because the technology failed, but because trust did.
If you are preparing a board-level AI case and want a second opinion on how it is framed, Rodan offers a structured diagnostic engagement - typically completed in four to six weeks - that validates the problem, stress-tests the commercial case and gives you the foundation for a proposal that holds up to scrutiny.
