AI is changing leadership: The new rules for Leadership 4.0

Why leaders no longer need to know all the answers and why orientation, responsibility, and human judgment are becoming more important?

For decades, the role of leadership was relatively clear. Leaders were expected to provide guidance, make decisions, and take responsibility. Those higher up in an organization generally had more experience, more information, and ideally, a better overview of where the company should be headed. Leadership was therefore also associated with an unspoken expectation: when in doubt, the leader knew the answer.

But this very understanding of leadership is currently under pressure. Not because leadership is becoming less important—quite the opposite. The conditions under which leadership takes place are fundamentally changing, and artificial intelligence is accelerating this development. It is not only changing individual processes or tools, but increasingly also the way knowledge is generated, decisions are prepared, and tasks are distributed. Activities that were considered standard practice for certain roles for years can suddenly be supported or partially taken over by AI. Employees can create analyses, prepare presentations, develop ideas, or have AI provide them with different perspectives on a problem within minutes. At the same time, systems are emerging that no longer simply react to individual prompts, but process tasks more independently and use a variety of tools to do so.

All of this is happening at a speed that traditional organizational structures can hardly keep up with. Job descriptions, responsibilities, and processes are changing more slowly than the technology that is currently challenging them. This puts managers in an unfamiliar situation: they are expected to provide guidance, even though they themselves cannot yet know exactly what work will look like in two or three years.

The crucial question, therefore, is not just how companies can successfully implement AI. The bigger question is: How do we lead people in a working world whose development we ourselves do not yet fully understand? This is precisely where Leadership 4.0 begins.

Leadership no longer means having an answer for everything.

Perhaps one of the biggest changes in modern leadership is being able to say a seemingly simple sentence: "I don't know yet."

In a traditional understanding of leadership, this statement can quickly be perceived as a weakness. After all, employees expect guidance from their leaders. They want to know what will happen next, what decisions will be made, and what changes will mean for their own work.

These expectations don't disappear in the age of AI. Leaders simply can no longer fulfill them with the same certainty as before. They are meant to provide people with security, yet they themselves are in the midst of a transformation whose outcome is still uncertain. They are not familiar with every new technology, they cannot reliably predict which tasks will be performed by humans and which by machines, and they often don't even know how the roles within their own teams will change.

The answer cannot be to feign certainty. Rather, modern leadership must learn to deal with uncertainty credibly. A leader may say that they don't yet know the definitive answer. At the same time, they must clearly communicate what is already clear, what direction the company is taking, and within what limits experimentation is possible.

Leadership 4.0 therefore does not mean less guidance. It means a different kind of guidance. Leaders no longer necessarily prescribe a finished path. They create a framework within which a team can develop this path together.

Current research on the introduction of agent-based AI points in a similar direction. McKinsey describes a key leadership task as providing clear direction, even under uncertainty, while simultaneously modeling the learning behavior expected of employees. This significantly changes the role of the leader. The person who distributes answers increasingly becomes someone who asks good questions, reveals connections, and ensures that viable solutions can emerge within the team.

Responsibility shifts to the team. Leadership does not disappear as a result.

If leaders can no longer dictate every course of action, the distribution of responsibility must inevitably change. This is precisely a second important shift of Leadership 4.0.

Leadership must break free from micromanagement more decisively than ever before. Employees need freedom, trust, and the opportunity to truly take on responsibility. This may sound like a familiar leadership principle, but AI gives it a new urgency. An organization can hardly unlock the potential of this technology if every idea, experiment, and decision has to navigate multiple levels of hierarchy.

Many useful applications don't originate in a central AI department, but rather where people are most familiar with their daily problems. In marketing, the team itself knows which coordination processes waste time. Sales is familiar with the recurring tasks involved in preparing for customer meetings. In customer service, employees know exactly which questions need to be answered every single day.

If these people are to use AI effectively, they need room to maneuver. However, this doesn't mean that everyone should simply try out any tool with any company data. Precisely because AI systems are becoming more powerful and autonomous, managers need to define the boundaries more clearly.

Which data may be used? Which results require human review? Which decisions may a system prepare, and which should it never make on its own? Where does the ultimate responsibility lie?

Leadership 4.0 therefore does not mean withdrawing from leadership. It means organizing control differently. Good leadership no longer tries to control every single step of the process. It creates clear guidelines and, within these guidelines, enables as much individual responsibility as possible. This is a crucial difference. Anyone who wants to delegate responsibility to a team must first clarify the scope that team is actually allowed to shape.

Those who want to lead change should have experienced it themselves.

Another consequence can be drawn, which is particularly important for managers: It is not enough to simply understand AI strategically. Anyone who wants to lead others through this change should have experienced firsthand what it is like to work with this technology.

This doesn't mean that every CEO has to learn programming or every marketing manager has to develop complex AI agents. Rather, it means taking a real problem from their own daily work and trying to solve it with AI.

Only at this moment does an abstract technology become a concrete experience. A leader then experiences firsthand how impressively well a system can function one moment and how surprisingly wrong it can be the next. They realize how much context is necessary for usable results, where an instruction was too imprecise, and at what point human oversight remains indispensable.

Leaders should build and experiment themselves before expecting their teams to adopt new ways of working. Those who have personally experienced the possibilities and limitations of AI can speak much more credibly about what it means for their own organization.

This is also relevant because employees closely observe how seriously a change from above is actually taken. If a manager describes AI as strategically crucial but has no personal experience with it, a contradiction arises. The unintended message then becomes: You should change the way you work, but mine will remain the same. Leadership 4.0 therefore doesn't require technical perfection. But it does require curiosity and a willingness to learn.

People are more likely to accept change if they can help shape it.

Many transformation projects still follow a familiar logic. A new system is selected, a rollout is planned, employees are informed and trained, and at some point, the change is officially considered implemented. With AI, this model quickly reaches its limits because it's not simply a matter of introducing a new tool. In some cases, the work itself changes. And thus, the change directly affects the people who have performed that work up until now.

Three questions that are crucial for successful change: How do we provide people with guidance? How do we turn those affected into participants? And how do we empower them to cope with the change?

The second point in particular deserves more attention. Employees shouldn't only encounter AI when a finished solution is presented to them. They should be involved as early as possible in identifying where this technology can truly be helpful. The starting point isn't the question of where the company still lacks an AI use case. A much more relevant question is where unnecessary friction arises in daily operations.

One of our first experiments involved using an AI-powered telephone assistant. The idea didn't stem from an abstract AI roadmap, but from a very concrete problem. Incoming calls interrupt focused work, and at the same time, they might be from an important customer that no one wants to miss. So, we tried to solve precisely this problem with AI. This example shows how the perception of AI can change. Employees then don't experience technology being forced into their work. They experience a problem that occupies them every day disappearing. Those affected become stakeholders. And those stakeholders can, over time, become multipliers.

This is precisely one of the most powerful opportunities for companies. The most convincing advocates for new ways of working are often not the members of a central transformation team, but colleagues who have tried something themselves and can say: This actually works, this helps me in my work.

Leadership 4.0 must also be able to talk about fear.

At this point, the discussion about AI quickly becomes uncomfortable. Because behind many questions about the introduction of new technologies lies a much more personal question: What does this mean for me?

This question cannot be answered with productivity indicators.

Let's take a marketing employee who has been creating presentations for many years. She knows the corporate design, understands what content the sales department needs, finds suitable images, and ensures that everything looks professional in the end. A large part of this work is part of her professional identity. She has become very good at it.

Now, an AI system can suddenly complete large parts of this task in just a few minutes.

From a company perspective, this can quickly be turned into a positive story: The employee gains time and can dedicate herself to more strategic tasks in the future. However, from her perspective, a completely different question may initially arise: If the machine now does what I've become good at over the years, what value do I still have?

The connection between task, competence, and identity should not be underestimated. When companies talk about AI, they are therefore not only talking about more efficient processes. They are also talking about people who may lose some of what they have previously used to define their value within the organization.

Leadership 4.0 requires more than a communication campaign in this situation. Leaders must take such concerns seriously and simultaneously develop a credible perspective on the new responsibilities that arise.

Because the work doesn't necessarily disappear. It just shifts its focus.

The more AI produces, the more important human judgment becomes.

Today, AI can quickly develop multiple campaign ideas, prepare a presentation, structure information from various sources, or formulate different strategic options. This shifts some of the human work from pure production to evaluation.

Carolina describes this development in conversation using the image of a judge role . People don't necessarily have to perform every single step of the process themselves. Instead, they need to be better able to assess which result is usable, which assumptions are wrong, and which decision fits their own organization.

This change initially sounds like a relief, but in reality it places higher demands on people.

Anyone who writes a text themselves bears direct responsibility for its content. Anyone who has an AI generate five texts for them must be able to decide which one is good. Anyone who automates an analysis must be able to recognize whether its conclusions are plausible. Anyone who has multiple strategies generated needs a sufficiently deep understanding of the business to be able to choose the right one.

The more options AI generates, the more important the ability to distinguish between them becomes.

This also changes the role of leadership. The crucial question is no longer simply how employees can complete certain tasks faster. Leaders must support people in making better decisions. They must develop judgment, contextual understanding, and a sense of responsibility.

The World Economic Forum arrives at a remarkably similar assessment in its Future of Jobs Report 2025. While skills related to AI and Big Data are gaining significant importance, human competencies such as analytical and creative thinking, resilience, leadership, and collaboration remain crucial.

The future of work therefore probably lies less in a competition between man and machine than in a new division of labor between the two.

Perhaps the wrong question is: How much faster will AI make us?

Many companies hope that AI could make people more efficient. I disagree with this idea on one crucial point. The goal isn't simply to make people faster and then have to handle even more projects. The real opportunity lies in becoming better because it frees up more time for reflection.

This shift in perspective is important because a large part of the current AI debate is dominated by productivity. Companies ask how many hours can be saved, how much more content a team can produce, or which processes can be accelerated even further. This is understandable, but it doesn't go far enough.

If AI merely ensures that an already overloaded work system runs even faster, the end result may only be more output: more presentations, more analyses, more emails, more decisions, and presumably, eventually, even more meetings.

The more strategically interesting question is therefore: What do we want to do with the time we have gained?

Perhaps the true value of AI doesn't lie in making people work ever faster. Perhaps it lies in finally allowing them to slow down certain things again. To prepare more thoroughly for a difficult customer conversation. To avoid making a decision between two appointments. To truly listen to a colleague. To fully develop a strategy before the next task appears on the screen.

That would be a different understanding of productivity. One that measures not only output, but also quality.

This idea is central to Leadership 4.0. Leaders play a crucial role in determining whether increased efficiency is simply filled with new tasks or whether it actually creates space for better work.

Change cannot simply be rolled out.

This is another reason why the traditional idea of a large-scale AI rollout is only partially suited to this new world of work. The technology is changing too rapidly, and many of the best use cases only emerge when people start working with it.

My personal approach is a wave approach . Instead of trying to change the entire organization immediately, you start with a small group of people who are curious and want to experiment. They develop initial applications, gather experience, and solve concrete problems. When others see that something works, the next wave emerges.

This approach is interesting because it takes into account how change often actually works in organizations. People don't adopt new behaviors simply because they are convinced by a presentation. They observe their colleagues. They see that someone solves a problem differently, that a new way of working functions, and that the feared disaster doesn't materialize.

This is how trust is built.

The first AI assistant, tested by three people, can eventually be used by ten. Those ten will become twenty. With each new application, not only does the technical expertise grow, but also the organization's ability to figure out for itself what works and what doesn't.

This also changes the understanding of transformation. AI is not a project that will be completed at a specific point in time.

The most important skill might not even be AI competence.

Models will improve. New applications will emerge. Agents will take over tasks that are currently performed manually. And many of the tools that companies are currently discussing intensely will either be commonplace or already replaced in a few years.

Companies cannot start from scratch with every new technological development. They need a capability that goes beyond individual tools: the ability to continuously adapt; this requires organizational competence.

Change should no longer be viewed solely as a project with a beginning, a roadmap, and an eventual end. Organizations must learn to continuously evaluate, test, and integrate new possibilities into their working methods.

This is a key concept for Leadership 4.0. A leader's task is not to prepare their organization for some future, stable end state. This end state will likely never exist. The task is to develop an organization that can learn as its environment changes.

This requires people who are allowed to take responsibility, leaders who don't have to pretend they already know everything, and a culture in which a failed experiment isn't automatically interpreted as a personal failure. In this environment, psychological safety isn't just a feel-good factor, but a prerequisite for learning to take place at all.

The more intelligent our machines become, the more human leadership must become.

We are developing systems that can process information faster, write texts faster, generate analyses faster, and perform more and more tasks independently. The obvious reaction to this would be to demand ever greater speed from people as well. Perhaps that would be precisely the wrong conclusion.

As machines take over an increasing share of production, the value of skills that cannot simply be improved by increasing speed also rises. These include listening, reflecting, understanding connections, tolerating uncertainty, weighing decisions, and taking responsibility.

Leadership 4.0 therefore does not mean that leadership has to become more technical. Leaders need an understanding of what AI can do and where its limits lie, but their fundamental task remains profoundly human.

They must provide guidance without feigning certainty. They must enable responsibility without relinquishing it. They must encourage people to try new things while simultaneously defining clear boundaries, and they must ensure that the time technology may give us back is not immediately filled with even more work.

Perhaps this is the most interesting consequence of this technological development: the more powerful artificial intelligence becomes, the clearer it becomes which tasks we should not delegate to machines.

The more intelligent our machines become, the more human leadership must become.