At our recent Responsible Futures roundtable, we brought together senior leaders from healthcare, education, insurance, pharmaceuticals, and technology to discuss a question that is becoming increasingly common across every sector: what is really holding organisations back from adopting AI?

Our community discussed governance, regulation, leadership, funding, technology, and organisational design. Yet despite the breadth of perspectives in the room, the discussion repeatedly returned to the same place: people.

Not in the sense that people resist change, a phrase that has become almost synonymous with digital transformation, but in the sense that successful change depends on understanding how people experience it. Again and again, participants described organisations that had invested in technology, identified valuable use cases and established governance frameworks, only to find that meaningful adoption remained frustratingly difficult.

What emerged was a different way of thinking about the problem. Perhaps the greatest barrier to AI adoption is not capability, regulation or even organisational inertia. Perhaps it is confidence.

______________________________________________________________________

For many organisations, the instinctive response to slow adoption is to invest in training. If people aren't using AI tools, the assumption is often that they need to become more capable. While technical capability is undoubtedly important, it overlooks a more fundamental question: what if people already feel highly capable?

Most employees have spent years, often decades, developing expertise in their profession. That expertise is rarely limited to formal qualifications or documented processes. It is built through experience, judgement, relationships, and thousands of small decisions made over the course of a career. It is knowing when to trust the data and when to challenge it. It is recognising patterns that cannot easily be written into a policy document. It is understanding the context surrounding a decision, not just the decision itself. This expertise becomes more than a skillset. It becomes part of a person's professional identity.

When organisations introduce AI, their intention is almost always to enhance this expertise rather than replace it. Yet that distinction is not always experienced by the people being asked to change. A new AI-enabled workflow can easily feel like an implicit judgement on the old one. The message leaders believe they are communicating: "there may be a more efficient way of doing this", can be interpreted as: "the way you've been working for the last twenty years is no longer good enough."

That is a very different conversation.

Perhaps this explains why resistance to AI is so often misunderstood. What appears to be reluctance towards technology may instead be uncertainty about professional value. People are not simply being asked to use a new tool; they are being asked to reconsider the methods that have made them successful throughout their careers.

______________________________________________________________________

In many roles, expertise is closely linked to autonomy. Senior professionals are trusted not because they follow a process perfectly, but because they know when experience should override the process. Their judgement is what makes them valuable. If AI is introduced in a way that feels prescriptive or controlling, it can be experienced as a reduction in professional autonomy rather than an enhancement of capability. Training alone cannot solve that problem.

This is not the first time organisations have faced a challenge like this. When businesses transitioned from paper records to digital systems, many of the same concerns emerged. The resistance was rarely about the technology itself. It was about changing established ways of working and asking experienced professionals to place their trust in unfamiliar systems. The organisations that navigated that transition most successfully recognised that digital transformation was not simply about implementing new technology; it was about helping people carry their expertise into a new way of working.

AI presents a similar challenge, but on a much larger scale.

Unlike many previous technologies, AI appears to engage directly with activities that have traditionally been considered uniquely human: analysing information, generating ideas, making recommendations, and supporting decision-making. As a result, the change feels inherently more personal. The conversation is no longer about learning a new piece of software; it is about redefining the relationship between human judgement and technology.

This may also explain why so many AI initiatives show promise in pilot programmes yet struggle to achieve organisation-wide adoption. Pilots are typically conducted with enthusiastic volunteers, supported by dedicated project teams and protected from many of the operational realities of day-to-day business. Scaling requires engaging the wider organisation, people who did not volunteer, who are already highly competent in their roles and who may reasonably ask why successful ways of working now need to change.

The challenge, therefore, is not simply to demonstrate that AI works. It is to create the conditions in which people feel confident enough to explore how it can improve the work they already do. That requires a different approach to leadership.

______________________________________________________________________

Successful organisations are unlikely to treat AI as a technology programme alone. They will recognise it as a change programme that happens to involve technology. They will involve people early, create opportunities for safe experimentation, and communicate clearly that AI is intended to complement human expertise rather than diminish it. Most importantly, they will acknowledge that asking people to change how they work is never just a procedural request; it is often a deeply personal one.

One of the most valuable outcomes of our Responsible Futures discussion was not a definitive answer, but a shared recognition that the conversation around AI needs to become more human. We spend considerable time discussing governance, regulation, investment, and capability, all of which are important. Yet if organisations overlook confidence; people's confidence in their expertise, their judgement and their place in the future, they risk solving the technical challenges while leaving the human ones untouched.

______________________________________________________________________

Ultimately, organisational transformation has always depended less on technology than on people's willingness to embrace it. That willingness cannot be mandated through policy or created through training alone. It is built through trust, respect and a genuine recognition that experience remains valuable, even as the tools we use continue to evolve.

Perhaps, then, the question organisations should be asking is not "How do we get people to adopt AI?" but "How do we help people feel confident that their expertise still matters in an AI-enabled world?"

The answer to that question may determine whether AI becomes another promising pilot or a genuine organisational transformation.