The human edge of AI: why transformation cannot be left to technology

In brief

• AI transformation needs ownership from the executive team. Laing O'Rourke rolled out Copilot to 7,500 people overnight, but changing how those people work takes much longer.

• At Revolut Trading, a central AI function sets the architecture and risk standards. AI owners in each business area stay accountable for commercial outcomes.

• Experimentation builds confidence. To scale, it needs to connect to business outcomes and sit on solid data foundations.

• As AI takes on repeatable work, judgement and critical thinking become more valuable. Organisations need to rethink how people at every level develop them.

• The time AI frees up should go back into human work, such as better guest experiences at Travelodge and safer decisions at Laing O'Rourke.

AI may be moving faster than previous waves of technology, but one of the biggest questions facing organisations is surprisingly familiar: how do you turn technological possibility into meaningful change?

At our Women in Transformation Summit, our second conversation brought together leaders from Carnival UK, Laing O’Rourke, Travelodge and Revolut to explore the human edge of AI transformation.

The discussion deliberately moved away from the technology itself. Instead, we explored what AI means for leadership, skills, organisational design and the work that humans will continue to do best.

What emerged was a challenge for executive teams. AI cannot sit neatly inside the technology function because the decisions being made now will reshape work across the whole enterprise.

AI transformation cannot belong to one function

Louise Morgan, Head of IT – Group at Laing O’Rourke captured the problem through her experience where technology can be deployed remarkably quickly. For example, the organisation recently made Copilot available to 7,500 people at the flick of a switch.

Changing how 7,500 people think about their jobs is a very different proposition, however.

“I can enact transformation really easily. What I can't do is overnight change how all of those people are thinking about how they go to work and the jobs that they do.”

That distinction matters. If AI is treated primarily as a technology rollout, organisations risk measuring progress through deployment rather than value. The harder questions concern how work changes, which processes should be redesigned, what skills people need and where human judgement remains essential.

Jo Phillips, Chief People Officer at Carnival UK, described the importance of a strong partnership between people and technology leadership. But the panel went further. AI transformation ultimately has to become an executive responsibility.

As Louise put it, somebody around the leadership table should be continually asking what the organisation can become today that it could not have become three years ago, and how it needs to be architected accordingly.

Yana Shkrebenkova brought another perspective from her experience leading Revolut Trading. She described how a central AI function can provide the architecture, infrastructure, models and risk standards, while AI owners within individual business areas remain accountable for applying those capabilities to real commercial priorities. It is a useful model for balancing enterprise-wide consistency with enough ownership inside the business to turn AI into meaningful outcomes.

Start with value, not AI

Another thread running through the conversation was the danger of experimentation becoming fragmented.

Organisations understandably want people to explore AI. Experimentation builds confidence and reveals opportunities that a central team might never identify. But Jo also described the tension between that bottom-up energy and a more joined-up enterprise view.

Yana also challenged organisations that are waiting for the AI landscape to settle before making bigger decisions. With the technology evolving so quickly, she argued that certainty may never arrive in the way leaders would like it to. The more practical response is to create safe environments where teams can test different tools and models, invest in education and build their understanding as the technology develops, rather than allowing uncertainty to become a reason for inaction.

Individual functions can bring forward compelling ideas while the organisation still misses bigger opportunities elsewhere. The answer is not to shut experimentation down. It is to connect it more deliberately to business outcomes.

That also requires foundations. Kinnari Ladha, Chief Data Officer at Travelodge, highlighted the less glamorous questions that determine whether AI can scale: is the data connected and accurate? Who can access it? Is the architecture adaptable enough for technologies that will continue to change? Do people have the literacy to use what is being introduced?

Without those foundations, organisations can accumulate isolated pilots without building the capability to turn them into enterprise value.

Human skills are becoming more valuable, not less

The conversation became particularly interesting when we explored what happens as AI takes on more rules-based and repeatable work.

The panel discussed judgement, empathy, accountability, curiosity and critical thinking, qualities historically dismissed as “soft skills”. In an AI-enabled organisation, they become much harder currency.

Jo argued that these capabilities cannot remain the preserve of senior leadership.

As more people work alongside AI, employees throughout the organisation will need to question recommendations, understand context and consider the human consequences of decisions. That raises an important challenge for early careers too. If AI removes some of the foundational work through which previous generations developed experience, organisations will need to think much more intentionally about how future leaders develop judgement.

This is not an argument for protecting repetitive work simply because humans used to do it. It is an argument for redesigning development alongside work.

Use AI to create more space for being human

Perhaps the most optimistic part of the discussion came when we turned the question around.

If AI can remove friction, what can people spend more time doing?

For Travelodge, that could mean giving 13,000 colleagues more time to create better experiences for 22 million guests. For Carnival UK, it means removing low-value work so people can focus more intensely on creating memorable human experiences. For Laing O’Rourke, it includes helping people make better and safer decisions in an industry facing significant skills shortages.

The human edge, then, is not about competing with AI at the things machines increasingly do well.

It is about being much more deliberate about the things we want people to do with the capacity AI creates.

That requires technology, data and governance, but it also requires organisations to rethink leadership, skills and work itself. AI transformation will succeed when those conversations happen together, rather than when one function is left to carry the change for everybody else.

Want to know more?

If the themes explored here resonate with you, we’d love to continue the conversation. Our Women in Transformation community brings together senior women working across change and transformation to share experiences, challenge thinking and learn from one another. If you’d like to find out more about the community, explore joining us at a future event or simply discuss any of the ideas raised in this article, get in touch with the S&S team.