When it comes to AI adoption, I keep noticing how unusually high the level of interest is. Even the announcement itself triggers questions and discussions. Employees actively seek out licences, workshops are well attended, and prompt tips and best practices are shared.
The excitement about suddenly possible ways to make work easier is real. But so is the uncertainty that comes with it: Can I still learn this? What does AI mean for my role? Will parts of my work disappear, or will my team look different in the future?
Excitement and uncertainty can exist at the same time. And that has consequences for transformation management.
In change projects, high participation is usually seen as a good sign. With AI, we often see even higher levels of engagement than in other initiatives. But participation does not automatically mean unconditional support.
Menschen können eine AI-Einführung interessant finden, aktiv mitmachen, ihren Nutzen sehen und sich trotzdem bestimmten Konsequenzen widersetzen.
This becomes particularly visible when identifying use cases. Someone who suspects that automating a process could eventually call their own role, or that of a team member, into question may have little interest in suggesting exactly that use case.
The same applies to productivity gains. Employees may use AI extensively and make their own work significantly more efficient without making every automation opportunity or every hour saved openly visible – particularly when it is unclear whether this might later lead to the redesign or downsizing of their team.
From an implementation perspective, this is hidden resistance. From an individual perspective, it is entirely rational. And it does not mean that the same person does not fundamentally support the introduction of AI.
This is what makes AI adoption particularly challenging for change and transformation management. Resistance does not always show up as open rejection. It can be selective – especially because the technology does not only enable new ways of working, but also changes tasks, roles and skill profiles. In some areas, it may also mean that certain activities or jobs disappear. This affects employees' personal interests very directly. And anyone familiar with change knows: this is where things get tricky.
Assessing AI adoption solely through participation rates, licence usage or workshop attendance is therefore misleading. These metrics show activity. They do not automatically show how people feel about the potential consequences of the change. The resulting picture may look more positive than people's actual willingness to support all the implications of the transformation.
Personal contact can bring us closer to what is really happening. Ambassadors or champions should therefore not be understood simply as multipliers. Their role is also to feed experiences from across the organisation back into the transformation: Where is uncertainty emerging? Which assumptions do not hold up in practice? Where are there obstacles that the project team cannot yet see?
This is not about using them to spy on sentiment. It is about identifying what is still not working and improving it. But that only works when there is a relationship of trust in which they can openly raise uncertainties, resistance and observations from their teams. Especially when these do not fit the positive adoption story.
It is equally important to make potential changes to roles visible early on. Not every future development can be predicted. But when it becomes clear that tasks will shift, skill requirements will change or teams will be reconfigured, people can be prepared for that: through conversations, transparent role development and targeted training.
Trust grows when organisations are clear about what they know, what they do not yet know and the criteria according to which decisions will be made. This is particularly important when jobs may actually disappear.
With AI in particular, it therefore helps not to think of support and resistance as opposites that cancel each other out. The key question is not only whether people participate. It is also what happens once an interesting new tool starts changing their own work.