The tools are genuinely impressive. The capability is real. And the temptation for founders, for leadership teams, for anyone who has spent time with what these tools can do is to start building. To expand. To create things that weren’t possible twelve months ago.
That’s not wrong. But it carries a risk.
AI multiplies your capacity to produce. It multiplies output, execution speed, the volume of things you can create and the pace at which you can create them. What it doesn’t do, what it has never done and won’t do, is tell you whether you’re solving a problem that actually exists and is worth solving.
That clarity has to come from the business.
The UK data on AI adoption is instructive here, though not in the way most people expect.
According to Deloitte’s 2026 research, only 6% of UK businesses have employees using AI tools daily. The majority are still earl exploring, piloting, working out where it fits. And yet 75% of those who have engaged with it are already seeing improved productivity.
So the technology works. That’s not the question.
The question is what you’re pointing it at.
Because a business that isn’t anchored to a clear commercial outcome doesn’t get more focused when it adopts AI. It gets more productive at going in the wrong direction. The same research notes that only 18% of UK businesses have seen increased revenue from AI so far, despite the efficiency gains. Productivity without commercial clarity doesn’t convert. That gap between efficiency and actual commercial return is telling you something. It suggests that a significant number of businesses are using AI to do more, without being sufficiently clear on whether the more they’re doing is the right more.
This is the thing I find myself coming back to in conversations with founders.
The businesses genuinely pulling ahead aren’t necessarily the ones who adopted AI first or who have the most tools in their stack. They’re the ones where the leadership team is clear about what problem they exist to solve for their customers, and what good looks like when they’ve solved it.
That sounds basic. It should be basic. But in a scaling business, with competing priorities, commercial clarity is harder to hold than it looks.
And AI, if you’re not careful, makes that harder. Not easier.
Because the technology is extraordinarily good at generating. Ideas, content, analysis, product features, processes, reports. The constraint is no longer execution capacity. The constraint is judgment. Knowing which of the twenty things you could build is the one that solves a real problem for a real customer who will pay for it.
That judgment is human. It has to live somewhere in your team. And if it doesn’t, if the business is running on momentum and output rather than on a clear picture of the outcome it’s delivering — AI will give you a problem, not resolve it.
What’s interesting is that the most forward-thinking organisations are already building their teams around this reality. And the capabilities they’re investing in tell you something important about where commercial advantage is going to sit.
The first is what some researchers are starting to call AI orchestration. It sounds technical. It isn’t. What it actually means is the ability to direct AI effectively, to be clear enough about the outcome you want that you can shape what the technology produces, judge whether what comes back is right, course-correct when it isn’t, and know when a decision needs a human rather than a model running in the background. This is a strategic skill, not a technical one. It requires clarity about the problem before anyone touches the tools. The people who are good at it tend to be the people who are genuinely clear about what they’re trying to achieve and why.
The second is data literacy and again, not in the way people usually mean it. This isn’t about being able to build dashboards or write code. It’s about being able to look at what AI surfaces and ask the right questions about it. To understand what the output is actually telling you, what it might be missing, and what decision it should or shouldn’t inform. When AI is generating analysis and insight at volume, the human contribution isn’t the analysis. It’s the quality of the questions being asked of it and the judgment about what to do with the answers.
The third is adaptability not as a soft skill on a CV, but as a genuine operating capability. The shape of work is changing faster than most businesses are updating their thinking about it. The ISER Essex research published earlier this year is clear that productivity gains from AI only materialise when businesses invest in process redesign alongside the technology. That means people who can work differently as the context changes and be open minded to different ways to approach a problem.
And underpinning all of it is outcomes focus. The shift happening in how the most effective organisations structure their teams is away from activity and toward accountability for results. McKinsey’s latest research describes teams built around workflows and outcomes, where AI handles the execution of repeatable tasks and humans own the judgment about whether the direction is right. In that model, being busy is not the same as being valuable. What matters is whether the work you’re doing is moving the business toward a specific, defined result for a specific customer.
This is where workforce planning comes in and it’s worth being precise about what that means, because it’s not primarily about headcount or hiring.
Workforce planning at its most useful is a commercial question. It asks: given where this business is trying to get to, and given what AI can now do, what do we actually need our people to be capable of? What judgment calls need to sit with a human? What relationships require a person? What decisions are complex enough, ambiguous enough, or high-stakes enough that they can’t be handed to a model?
Most scaling businesses haven’t asked those questions yet. They’ve grown by adding people to meet demand, by promoting from within when something needed covering, by reacting to gaps rather than designing around a clear picture of what the business needs to be capable of. That approach worked well enough when execution was the constraint. When AI removes that constraint, it stops being sufficient.
The businesses that get this right aren’t thinking about it as an AI problem or a technology problem. They’re thinking about it as a commercial clarity problem. They’re getting specific about the outcome they’re delivering for customers, and then working backwards to ask what that means for the capabilities their team needs.
The McKinsey State of Organisations report published this month talks about “AI-first by design” structures, teams built around outcomes and workflows, with AI handling the execution of repeatable tasks and humans owning the judgment about direction. The framing that works for me is this: the human layer isn’t about doing more. It’s about being clear on what problem is needs solving and is worth solving before anyone starts doing anything i.e the why?
Most businesses have never had to articulate that explicitly. The pace of work meant you could get away without it. When capacity is the constraint, you hire, you execute, you grow. The doing was enough to validate the direction.
When AI removes the capacity constraint, the direction is what’s left. And if the direction isn’t grounded in a genuine customer problem, one that actually exists, that someone will actually pay to have solved, you find out the expensive way.
And this is the part that often gets missed in conversations about AI and business strategy. The commercial clarity question isn’t just a strategic exercise. It has to be carried by the people around you. Does your leadership team have the capability, the judgment, and the clarity to lead and develop a team fit for the new reality. Do you have a plan for developing & hiring for the skills & capabilities e.g. AI orchestration, data literacy, genuine adaptability, outcomes focus in enough depth across your team to make the strategy executable?
A clear direction without the right people to make it happen is useless. The two have to move together.
The question I’d put to any founder thinking seriously about how AI fits into their business isn’t about technology at all.
It’s this: do you and your leadership team have a shared, clear, honest picture of the problem you exist to solve and what success looks like for the customer you solve it for?
If yes, AI is an extraordinary multiplier. You can move faster, build more, test more, deliver more, all anchored to something real.
If not, the priority isn’t the tools. It’s the clarity. And that’s a people and strategy problem before it’s a technology one.