We know that there is no shortage of big claims about AI.
Economic growth. Productivity gains. Transformed industries. Smarter services. New opportunities. Fewer inefficiencies.
The final panel session I attended at the Institute for the Future of Work‘s conference on Making the Future Work bought many of these themes together. The headline figures were certainly eye-catching: AI could add as much as £40 billion to UK GDP by 2030, and around three quarters of companies surveyed say they are already seeing productivity improvements from adoption.
That sounds impressive. And it is.
But the conversation was not just about the scale of the opportunity – it was the repeated reminder that AI adoption is not the same thing as successful transformation.
And that distinction matters. Because right now, much of the AI story is still concentrated in relatively narrow areas. Marketing and administration are among the most common functions where AI is being used, which tells us two things. First, adoption is happening. Second, it is still patchy. The benefits are not yet flowing evenly across sectors, workplaces, or regions.
This is important because when people talk about AI as a national productivity solution, it is easy to assume that adoption will somehow spread naturally and that value will automatically follow. In reality, the barriers are much more familiar: regulation, access to data, access to finance, organisational inertia, and the simple difficulty of changing how work gets done.
In other words, the challenge is not just technological. It is operational, structural, and human.
That came through clearly in some of the examples shared. One of the most interesting was healthcare. AI tools are being used in the NHS to support diagnostics, including lung cancer detection, while ambient voice technologies are helping reduce admin by capturing clinical notes in the background. Yet the point made was not that AI is replacing clinicians. In fact, in some cases it is increasing demand for skilled professionals, because better detection and faster insight generate more need for expert interpretation and treatment.
That is a useful corrective to some of the loud narratives around job displacement.
The message here was not that AI is removing work wholesale, but that it is reshaping work. It changes the mix of tasks, the pace of workflows, and the kinds of expertise that become more valuable. It can take friction out of the system, but it can also expose where systems, processes, or skills are not ready.
Which brings us to workforce transformation.
The UK’s ambition to up-skill 10 million workers by 2030 through the AI Skills Boost programme is a significant one, and the fact that more than 1 million courses have already been delivered shows that this is moving beyond just rhetoric. But the conversations also made clear that skills cannot be treated as a side issue. If AI changes work, then skills policy, workforce planning, and management capability have to move with it.
And that means thinking about inclusion as well as scale.
The Women in Tech Task Force was highlighted as an important example of this. If AI is going to shape the future economy, then the people designing, deploying, and governing it need to reflect society more broadly. Otherwise, we risk reproducing old inequalities inside new systems.
The same is true geographically. AI adoption could easily become another force that strengthens existing regional divides if investment, infrastructure, and innovation remain concentrated in the same places. That is why local examples matter.
They show that AI can be embedded in communities in ways that are practical, grounded, and relevant to local needs, rather than being treated as something that only happens in major tech hubs or policy circles. Which is a real takeaway.
The success of AI should not be measured only by how widely it is adopted, or how many tools are introduced, or even how much productivity improves in a narrow sense. It should also be measured by whether it helps people do better work, whether it opens up opportunity more broadly, and whether the gains are being shared fairly.
The clear message was that AI adoption is not the goal. Better work is.
And if we keep that in mind, then workforce transformation becomes about much more than technology. It becomes about designing systems, skills, leadership, and local ecosystems that allow AI to improve working lives rather than simply accelerate them.
