If you are a senior leader being asked to approve AI spending right now, you are hearing the word "transformation" a great deal. Vendor presentations are polished, the demos are compelling, and the pressure to move is very real.
The AI market is producing genuinely amazing and capable tools at an extraordinary pace, and the instinct to act is sound. The challenge is not whether to engage with AI, but how to evaluate what is being offered with the same rigour that applies to every other consequential investment.
The speed of AI advancement actually amplifies this challenge.
Something I have observed repeatedly in working with boards and executives on AI strategy is that the evaluation conversation tends to start with a question that assumes there is one destination. "Which AI product is best?"
It is a natural question, and mirrors how we evaluate most technology. But AI is different in two fundamental ways: it is better thought of as a capability rather than a technology, and AI doesn't carry inherent value in the way that framing implies.
AI value exists in the space between the tool's capabilities and a specific organisation's strategic objectives, wicked problems, real people, and capacity to grow and measure outcomes.
An AI platform that transforms one organisation can sit unused in another, and the difference is rarely related to the platform itself.
This is worth sitting with for a moment, I want to reframe the entire vendor conversation. If value is relational, then the productive question may be "is this product right for us?" It's a subtle shift that changes the frame to focus on how you will realise value (and away from a platforms novelty or amazing AI capabilities).
Taking this further (and only if this is a bigger investment envelope) consider these questions.
1) "What is the measurable problem or strategic objective AI help address?"
Put plainly, why? Everyone can say "efficiency" or some similar broad category, this is an invitation to clearly articulate the strategic relevance.
2) "If this is successful, which metrics will move by how much?"
This is the most common CFO question, so much so I now call it the "CFO test". Metrics also don't limit what success could target, as they could relate to staff satisfaction or revenue or NPS or customer onboarding times or error reductions.
3) "If Jane was great at AI but left us, how much of the AI benefits would likely leave with Jane?"
The third question relates to the 'adoption' metric, and specifically if it is underpinning promises of strategic capabilities.
This is framed to explore the balance between individual and cultural benefits with the expectation of investment into the organisations' strategic capabilities.
It is also interesting to explore how vendor's own behaviours maps to these questions. Is their commercial model structurally aligned to your objective? For example, if a per-user licence model creates an incentive to provide hyper-individualised features, what is the risk to efficient and integrated core processes required for scalable growth? Is the approach towards AI platforms seeking to advance systems thinking or 'a million different ways to do the same thing'?.
Lean into vendors who, before pitching their product, asks thoughtful and careful questions about the objective you're trying to achieve or problem you are trying to solve.
Lean into vendors who help build a measurement framework before the engagement begins or AI platform is procured.
That willingness to engage with the harder, less glamorous questions is itself one of the most reliable quality signals available to you. A vendor who insists on understanding your context before proposing their solution is telling you something important about how they work. And perhaps their motivation for a longer-term relationship built on mutual longer-term success.
While this principle isn't new, our hope is this will tell you a lot about the AI Platform, vendors, and help you navigate the wall of noise out there.