I have worked with technology for 35 years and with people for 15. In that time I have seen many waves: the internet, mobile phones, the cloud, social media. None of them brought so much unrest into teams so quickly as AI. That isn’t down to the technology. It is because AI starts exactly where we define ourselves most: in our thinking, writing and deciding.
If you run a company or lead a team, you currently face a double task. You are supposed to introduce AI sensibly. And you are supposed to make sure people don’t get left behind in the process. The first part is talked about a lot. The second far too little.
What the numbers show: relief and pressure at the same time
For its study “Arbeiten 2025” (Working 2025), the health insurer Pronova BKK surveyed around 1,230 employees in Germany in autumn 2025. The results, published in January 2026, are contradictory, and that is exactly what makes them interesting. Four in ten employees use AI tools regularly. By their own account they gain about six hours per week. More than half therefore leave work on time more often.
At the same time, two in three users say that the effort for checking and correcting has increased. More than half find their work more error-prone or less certain. Almost half fear that AI could make their own job redundant. Among those under 30, two in three feel put under pressure by the tools.
The insurer speaks of techno-stress and warns against immediately filling the time gained with new tasks. That matches what I experience in conversations. AI takes work off your plate, but it doesn’t take responsibility off your shoulders. And it raises the speed at which things are expected of you.
Why tool training alone is not enough
Most companies respond to the AI shift with training: How do I write a good prompt? Which tools are allowed? That makes sense, but it only answers the question of ability. The people in the room often ask a different one: What does this mean for me?
An employee who lies awake at night because they see their job at risk won’t calm down through a prompt workshop. A team leader who constantly has to check whether the AI results are right won’t be relieved if nobody clarifies their new role. When these questions stay open, you get what I used to see in IT with every major changeover: resistance disguised as a factual argument, and withdrawal that looks like disinterest.
Understand your own operating system first
I like to use an image from my IT days for this. Every person has an inner operating system: habits, beliefs, a nervous system that reacts to threat or safety. All programs run on this operating system, including the way you deal with AI. If the operating system runs under constant stress, all applications run worse. No software update changes that.
Under stress our view narrows. We reach for familiar patterns, check less thoroughly and decide faster from the gut. That is exactly what is risky with AI, because the results sound convincing even when they are wrong. A calm, clear head is therefore not a side issue. It is the precondition for checking AI results critically and making good decisions.
That is the idea behind my approach: understand your own operating system first. Then work with AI. It isn’t about slowing AI down. It is about people being able to work with it on an equal footing.
What resilience in a team means in practice
Resilience doesn’t mean enduring everything. It means coming back into your own strength after strain while staying able to act. In the AI shift it shows in a few very practical points:
- Clarity about roles: Who checks, who decides, who is accountable for the result? Uncertainty is a silent stress factor.
- Open handling of worries: When fear for one’s job is allowed to be a topic, it doesn’t disappear, but it no longer has to work in secret.
- Conscious use of time gained: Not every saved hour has to be refilled immediately. Recovery is part of performance.
- Everyday regulation: Short breaks, breathing exercises, a conscious switch between tension and calm. These are simple tools anyone can learn.
- Learning without blame: Mistakes in dealing with AI happen. Teams that can talk about them learn faster.
What executives can do now
You don’t have to set up a programme for this. Three steps are enough to start:
- Ask before you explain. What unsettles you most about AI? Listen without immediately reassuring or refuting.
- Make the ground rules visible. What may AI be used for, what not, and who is ultimately responsible? In writing and short.
- Give room for regulation. Deliberately plan breaks and simple body exercises into workshops and meetings. That often works better than another slide.
And one more thing: you are a role model, whether you want to be or not. If you seem rushed yourself, the team will adopt that. If you radiate calm and honestly say what you don’t know yet, that gives security.
A bridge between technology and people
I notice how much of what we know about resilience today is old human experience. Breathwork, meditation and trance have existed in many cultures for thousands of years. Modern stress research confirms much of it: slowing your breath calms the nervous system. Learning in silence to observe thoughts creates distance. That isn’t esoteric, it is tried-and-tested craft. And it fits remarkably well in a time when machines keep getting faster.
The more technology takes off our hands, the more important becomes what only humans can do: judgement, compassion, responsibility and the ability to stay calm in unrest. These abilities can be practised.
How I support companies with this
I support executives and teams in and around Düsseldorf and online. That can be one-to-one coaching for an executive, a seminar for a team or a day format on resilience and dealing with AI. I combine my 35 years of IT experience with my work as a psychological counsellor and stress expert. You’ll find more on the Services page and under Coaching in Düsseldorf.
If you’d like to know what suits your team, write to me or book a no-obligation intro call.
Source: Pronova BKK, study “Arbeiten 2025”, press release of 26 Jan 2026 (press release, in German).