The Question Every Board Should Ask About AI (But Doesn't)

In the past two years, almost every board I have spoken with has asked some version of the same question: “Should we be investing in AI?”

It is a reasonable question. It is also the wrong one.

The right question - the one that almost no board asks - is this: “What decisions are we currently making about AI, who is advising us on those decisions, and are they qualified and unconflicted enough to advise us well?”

This distinction matters more than most directors realise.

The Gap Between Governance and Technology

Boards are generally good at governing things they understand. Financial performance, regulatory compliance, risk management, executive succession - these are domains where directors have pattern recognition built on decades of experience. They know what good looks like. They know which questions to ask. They know when they are being managed rather than informed.

Technology is different. Not because directors are incapable of understanding it, but because the pace of change has outstripped the cycle time of board-level learning. Generative AI went from research curiosity to boardroom agenda item in approximately eighteen months. Most directors are still building their mental models of what it does, let alone what it means for their organisation’s strategy, risk profile, and competitive position.

This creates a governance gap. The board is responsible for overseeing decisions about AI - investment, risk, ethics, workforce impact - but may lack the independent expertise to evaluate whether the advice it is receiving is sound.

The Advisor Incentive Problem

This gap is typically filled by one of three sources: the organisation’s internal technology leadership, an external consulting firm, or a technology vendor. Each has structural limitations.

Internal technology leaders are competent and well-intentioned, but they operate within the organisation’s existing assumptions. They may also have career incentives tied to AI adoption - the CTO who recommends caution is rarely celebrated as boldly as the one who launches a transformation programme.

Consulting firms provide frameworks and benchmarks, but their business model depends on large-scale engagements. A consulting firm that recommends a small, focused pilot - or no AI investment at all - is recommending itself out of revenue. The incentive is toward comprehensiveness, not precision.

Technology vendors provide technical depth, but their advice is inherently bounded by what they sell. A vendor assessment of your AI readiness will always find readiness for the vendor’s products.

None of these sources are dishonest. But none of them are unconflicted. And at board level, the difference between advice and sales is the difference between governance and theatre.

What Boards Actually Need

What boards need is not more information about AI. There is no shortage of AI briefings, white papers, and conference presentations. What boards need is the ability to evaluate the quality of advice they are receiving.

This requires several things.

First, it requires an independent perspective. Someone who does not sell technology, does not depend on ongoing consulting engagements, and does not report to the executives whose programmes are being evaluated. Independence is not a luxury in technology governance. It is a prerequisite.

Second, it requires translation, not education. Directors do not need to understand transformer architectures or fine-tuning methodologies. They need to understand the strategic implications: where AI creates genuine competitive advantage, where it introduces risk, where the organisation is and is not ready, and what the realistic timeline and cost look like versus what is being presented.

Third, it requires candour. Boards need an advisor who will say “this initiative is poorly conceived” or “your organisation is not ready for this” or “the vendor you have selected has a conflict of interest.” These are uncomfortable statements. They are also precisely the statements that protect shareholder value.

The Fiduciary Dimension

There is a fiduciary argument here that boards should take seriously. Directors have a duty of care that extends to technology decisions, particularly when those decisions involve material capital allocation, workforce restructuring, or the handling of sensitive data.

Relying solely on conflicted advice for material technology decisions is a governance risk. It may not create legal liability in every case, but it creates the conditions for poor decisions - decisions made on incomplete information, filtered through incentives that do not align with the organisation’s interests.

The standard boards apply to financial advice - independence, qualification, absence of conflict - should apply equally to technology advice. The sums involved are often comparable. The strategic consequences can be larger.

The Question to Ask

The next time AI appears on a board agenda, the most useful intervention a director can make is not to ask about the technology. It is to ask about the advice.

Who prepared this recommendation? What are their incentives? Do they benefit financially from the outcome they are recommending? Have we sought an independent perspective? Would we accept this standard of independence for financial or legal advice?

These are governance questions, not technology questions. And they are the questions that separate boards that oversee AI decisions from boards that merely approve them.

The technology will continue to evolve. The governance principles are not new. Boards already know how to demand independence, rigour, and candour from their advisors. The task is simply to apply those same standards to technology - before the decisions are made, not after the investment has been committed.