There is a question that is sitting on most boards' agendas right now, and it is not being answered well. The question is: what should we actually be doing about AI? And the reason it is not being answered well is that the organisations most commonly asked to answer it have a structural conflict of interest that prevents them from giving genuinely independent advice.
The conflict is structural, not personal
Most organisations rely on a managed service provider, a value-added reseller, or a cloud-aligned technology partner for their IT advisory. These organisations are not dishonest. Many of them employ highly capable people who understand their clients' technology estates in considerable depth. But they have a problem that is structural rather than personal: their revenue depends on selling products and platforms, and the advice they give is therefore shaped — inevitably — by the commercial relationships that sustain their business.
When a Microsoft partner advises a client on AI strategy, they are not lying when they recommend Microsoft Copilot. They are giving advice that reflects both their genuine belief in the product and their commercial interest in its deployment. The conflict is not visible on the surface. It is embedded in the incentive structure of the relationship. And it is present whether or not the adviser is aware of it.
This matters because AI adoption is not primarily a technical decision. It is a strategic one. The question of which AI tools to deploy, in which sequence, against which business problems, with what governance framework and what change management approach — these are leadership decisions, not technology decisions. And leadership decisions require advisers whose only obligation is to the quality of the advice.
The misdiagnosis has consequences
The technology industry's misdiagnosis of AI adoption as a technical challenge has had observable consequences. Copilot licences are sitting unused across thousands of UK organisations, not because the software is deficient, but because the deployment was not accompanied by the change management, the use-case definition, or the leadership alignment required to drive adoption. AI strategies that were written by vendors are being executed without ever having been properly evaluated for organisational feasibility. And boards are being given confidence assessments — "you are on the AI journey," "your AI maturity is at stage two," "you are ahead of your sector peers" — that are designed to reassure rather than inform.
The result is a market in which a significant proportion of AI investment is being made on the basis of advice that was never truly independent, against a strategic backdrop that was never honestly assessed.
What honest advice looks like
Honest advice about AI begins with a question that independent advisers are much more willing to ask than vendor advisers: is this organisation actually ready for what it is proposing?
Readiness is not a question about technology. It is a question about leadership clarity, data foundation, workforce capability, governance posture, and use-case discipline. Most organisations that have answered it honestly have discovered that their AI ambitions are ahead of their organisational readiness — which is not a reason to abandon those ambitions, but is an excellent reason to sequence the work differently.
Honest advice is also willing to say things that vendor advisers find commercially uncomfortable. That the proposed AI platform is the wrong choice for this organisation's specific situation. That the governance framework does not yet exist to deploy AI responsibly. That the highest-value use case is not the one the vendor is promoting, but the one that requires changes to the technology stack the vendor is selling.
What boards should do
A board that wants genuine independent advice on AI has three practical options. It can hire a senior AI adviser permanently — expensive, often unnecessary for organisations at the SME scale. It can engage a large management consultancy — disproportionately resourced and priced for enterprise, with their own incentive structures to navigate. Or it can work with an independent practitioner who has no products to sell, no platforms to promote, and no commercial interest in any particular answer.
The last option is relatively new in the UK market and still undersupplied. But it is the one that is consistent with the way that serious boards approach other areas where they need genuinely independent expert counsel — legal advice, audit, actuarial assessment. In those disciplines, independence is not a differentiator. It is the precondition of trust.
AI is now at the same point. The boards that understand this first will make better decisions. The boards that continue to rely on vendor-aligned advisers will continue to get vendor-aligned answers.
About the author
Stephen Harley is the founder of Requisite Intelligence and a former Field CTO at Bechtle UK. He holds an MBA with Distinction from Warwick Business School, with a dissertation focused on GenAI readiness in UK SMEs.