Walking through Davos this year, one thing was impossible to ignore. Not the scale of ambition. Not the number of conversations about AI. But the feeling underneath it all. A quiet tension.
On one hand, we are standing at the edge of what many describe as the next industrial revolution. On the other, we are navigating one of the most geopolitically volatile moments in recent history. Opportunity and uncertainty, present at the same time, in every conversation. Compared to previous years, something has clearly shifted.
For the first time in years, the tone around AI has clearly changed. There was less hype, fewer bold and sweeping visions. More grounded conversations about what has actually been done and what has not.
Leading companies were sharing concrete examples of scaling AI within a single function or clearly defined domain. They talked about results, governance, and what it takes to make things work in practice. A common pattern emerged: the organisations making real progress had invested heavily in rigorous governance models and controlled implementation. Structured execution, not experimentation for its own sake. And at the same time, something else was equally clear: no one has fully cracked enterprise-wide transformation yet.
One of the reasons Davos is so valuable is that it allows you to calibrate where the market truly is. Through hundreds of conversations, listening not just to what is being said, but how it is being said. Observing the level of clarity, confidence, and maturity behind the words.
This year, I focused specifically on how enterprise leaders think about AI. The shift was visible. Leaders are beginning to understand that AI is a transformation at scale, and the complexity of it is becoming real. But with that understanding comes something unexpected: a loss of confidence.
The narrative is maturing, but the path forward is far from clear.
One exchange captures the week better than any formal session. It happened in a corridor between sessions, the kind of brief exchange that happens when two people have thirty seconds and no reason to perform. I found myself next to a senior executive from a large European company, someone with real responsibility for their organisation's AI direction.
We talked for five minutes. What struck me was the register of the conversation. There was no confident narrative. No polished talking points. Just honest thinking out loud from someone genuinely grappling with the scale of what lies ahead. That informal register told me more than the formal agenda ever could. As the week went on, I kept noticing the same pattern: the more senior the person, the more honest about the complexity. The more they understood, the less certain they were about the path forward. That is a sign of maturity.
The most important shift I observed: AI is no longer a technology topic. It is becoming an operating model question. For years, the focus has been on embedding AI into existing structures, tools, processes, workflows. That approach is reaching its limits. Leaders are realising that AI doesn't just optimise the current way of working. It challenges it entirely and requires rethinking how decisions are made, how roles are defined, how work flows across the organisation, and how accountability is structured.
This is about redesigning the system itself, and that redesign must account for something that came up in almost every conversation: the rise of AI agents. The idea that AI can initiate actions, make decisions, and operate across workflows. The shift is real, and it is happening fast. But the understanding of what it actually takes to implement this well is still very limited. In particular, one critical aspect is consistently underestimated: the role of humans in the loop.
Many organisations assume that more autonomy automatically leads to more value. Without carefully designed interaction between human judgment and machine execution, the opposite can happen. Quality becomes inconsistent, decisions lack context, outputs become generic, and the organisation's DNA starts to erode. In the long run, that is a strategic risk.
There is excitement about the productivity potential AI enables, but also a growing awareness that the transformation required is far more complex than anticipated. There is, quite simply, no clear playbook yet.
One of the most striking things I heard during the week came from Jason Droege, CEO of Scale AI. He noted that successful AI transformation will require far more human consulting and transformation work than anyone initially expected.
This resonates strongly with what we see every day at Hellon. Off-the-shelf AI will not transform enterprises. Technology alone is not enough. The real challenge lies in redesigning how people work, redefining roles and responsibilities, building new capabilities, maintaining coherence and direction, and ensuring that organisations do not lose what makes them unique. AI amplifies, but humans define the outcome. They define quality, meaning and identity.
What emerges from Davos is a shift in focus. Leaders are beginning to understand that success with AI requires moving from pilots to scalable systems, designing operating models rather than just deploying tools, embedding governance and trust into core operations, and investing in human capability alongside technology. Most importantly, it requires accepting that this is not a linear transformation. It is a systemic one.
Despite the tension, or perhaps because of it, there is also a strong sense of opportunity. The next five years will be defining. Not just for how organisations adopt AI, but for how they evolve as systems.
At Hellon, we have believed since 2012 that business is ultimately human to human. That conviction has only grown stronger. Organisations are systems made of people, and the most impactful transformation, the kind that actually moves business outcomes, always happens people-first.
That belief brings clarity to everything we do: the quality of processes, the design of AI-enabled operating models, the changing role and identity of people inside organisations, and what it truly means to build an AI-native company that doesn’t lose its soul in the process.
That is the winning recipe. Human first, with AI as the amplifier.
If you’re wondering whether your strategy has this clarity, or whether your transformation is moving fast enough to capture the value ahead — let’s talk.
Now is the right time.
Jaakko Wäänänen is CEO of Hellon, a human-centred transformation consultancy with offices in Helsinki, Oulu, and London. hellon.com