Leadership reveals itself when no one is officially in charge. In many of the companies I work with, teams organise themselves. Since last year, tools have begun to take on more tasks. Hierarchies step back. And still, someone needs to set direction, build trust, hold focus and make progress possible. That is why I find The 21 Irrefutable Laws of Leadership by John C. Maxwell worth revisiting. Not as a formula for success, but as a framework for shared leadership. Especially in self-led teams and human–agent teams. What Maxwell calls “laws” are not about authority, they describe what becomes visible when leadership is distributed. And they show up in moments we often overlook:

when decisions are delayed
when misalignment goes unnoticed
when momentum fades and no one asks why

Human–agent teams challenge the old defaults. We do not lead tools the way we lead people. And yet we rely on their output. That gap needs structure. Leadership becomes framing, making assumptions visible and setting ground rules. Keeping humans in the lead (not only loop) without slowing them down.

In self-led teams, these are the friction points. Where roles blur or where responsibility is not given but must be claimed. With AI in the loop, the questions shift. What counts as collaboration? What belongs to the system, and what stays with the human? Who decides when the system speaks Maxwell’s strength lies here. He does not say who should lead. He shows what leadership requires. So we can recognise it better, practise it and build for it.

This series explores each law, one by one. I use the laws not only as inspiration for the leaders I work with and my own leadership practice, but also as a structure for thinking and asking the right questions. In the evolving relationship between humans and intelligent systems, leadership does not disappear, it moves and that is what makes it so challenging.

Daniel was a capable leader. Smart, experienced, known for running a tight operation. His team worked well together, especially in the office. Ideas flowed easily at the whiteboard, problems were solved over coffee, and there was a natural rhythm to how things got done. Then came 2020. Suddenly, everything shifted. Remote work became the norm. Uncertainty grew. The usual routines no longer applied.

Daniel responded the way many leaders did. He worked harder, extended his hours, added meetings, and started answering every message as quickly as possible. He wanted to stay close to the team and show that he was present. But the effect was the opposite. His team felt monitored rather than supported. Motivation dropped, decisions took longer, frustration quietly built up.

One night, an email arrived in his inbox: “It feels like you don’t trust us.” That sentence stopped him. He read it again, more slowly this time. And suddenly he saw the pattern. He was not helping. He was getting in the way. The more he tried to stay involved, the more he became the problem. So he changed. He reduced the check-ins to what truly mattered. He stopped handing out tasks and instead gave ownership. He stepped out of the details and focused on what the team needed to move forward on their own terms.

The shift was noticeable. His team stepped up, energy returned, and decisions started flowing again. They did not just survive the disruption. They found their footing and moved through it with confidence.

What about self-led teams?
The law still applies, just in a different form. When no one holds formal authority, it is the team’s shared leadership that sets the ceiling. If no one names the priorities, if communication stays vague or decisions are avoided, progress slows down. Even without hierarchy, someone has to lead the work. Otherwise the team loses momentum not because it lacks talent, but because it lacks focus.

And what about human–agent teams?
The law shows not only how people lead each other, but in how they organise the work around tools that do parts of the job. Tasks may be automated through AI Agents, but direction, standards and judgment are not. Someone still needs to define what matters, recognise when something goes wrong, and make sure the system fits the purpose.

When no one takes that role, tools start to fill the silence and even behave in unethical ways. Outputs go unchecked, edges get blurry. The team moves, but not always in the right direction. What looks like efficiency becomes noise and chaos. Not because the tools are bad, but because the leadership around them is missing. The lid is not (only) the technology, it is how we lead in its presence.

Questions I come back to
Do I trust the team, or do I stay too close?
Do we take shared leadership seriously, or wait for someone else to carry it?

When tools do part of the work, who makes sure they support the goal?
If things slow down or go off track, am I helping, or am I in the way?

Emma, a 42-year-old Head of AI at a global manufacturing company based in Munich, had worked hard for her promotion. She had put in nights and hours, delivered results, and finally stepped into a leadership role. She had the title, the raise and now the responsibility. But something was missing. Her team stayed distant. The connection was not there, and many meetings were quiet. People nodded but offered little and decisions dragged. And when she asked for support on something important, she met hesitation instead of commitment.

Frustrated, she turned to a mentor.
„Why won’t they listen? I’m their manager now.“
The mentor smiled.
„A title doesn’t make you a leader. Influence does. Right now, they’re following your position, not you.“

Emma realised she had leaned on authority instead of connection. So she shifted. She started listening more. She invited input instead of giving instructions. She showed up not to manage the work, but to support the people doing it. The change was gradual and people began to speak up. Share ideas. Take ownership. And one afternoon, she heard a teammate say, „Emma’s the kind of leader you actually want to work for.“ That was the moment she knew she had earned something a title alone could never guarantee.

What about self-led teams?
Influence matters even more where no one holds formal authority. In these teams, leadership emerges through action. Who speaks with clarity. Who takes responsibility. Who earns trust. The title may be missing, but the impact is felt. And others follow not because they have to, but because they choose to.

And what about human–agent teams?
Here, the idea of influence shifts again. Not toward the system, but around it. Tools do not respond to charisma. But people still do. Influence shows in how a team decides what to delegate, what to question, and what to double-check. It shows in how clearly someone frames a task, how confidently they speak for quality, and how responsibly they handle outcomes, especially when part of the work is handled by Ai-agents. In these teams, authority cannot be handed over to the tool. It stays with the human (I love to call it Human-in-the-lead). Leadership means making sure the system supports the work without replacing judgement. And that influence is earned by clarity, not control.

Questions to challenge your thoughts:
Do people follow because they trust me, or because they have to?
Am I leading through connection, or leaning on position?
In teams shaped by tools, who leads the choices we still have to make?

Imagine signing up for a marathon without training. You just show up on race day, expecting to run 42 kilometres. Sounds unrealistic. And yet, many people treat leadership the same way. One promotion, one course, one big moment, and they believe they are ready to lead.

Paul was one of them. Smart, driven, the person everyone turned to when problems needed fixing. He assumed leadership would come naturally. But after a few months, that belief started to slip. His team was distant. Some resisted his ideas. Others avoided him when issues came up. And when things went wrong, he felt the pressure to know what to do, even when he did not.

After one difficult meeting, his manager pulled him aside. „Paul, leadership is not a switch you flip. It is a process. The best leaders grow a little every day.“ This made Paul think, he stopped trying to prove himself and started to learn. He asked more questions. He watched how others handled pressure. He asked his team what they needed instead of assuming he already knew. He became more patient, less reactive and more open.

Over time, things began to change. His team grew more confident. They trusted him, not because he had all the answers, but because he had become someone worth trusting. When a big problem came up, no one waited for him to fix it. They stepped in on their own. That was the moment Paul understood. Leadership is not something you earn once. It is something you build, day by day.

What about self-led teams?
These teams do not work just because no one is in charge. They work when leadership behaviours are part of the culture. When people take initiative. When they speak up early. When they reflect and improve together. If a team avoids the hard parts of leadership, like giving feedback or making tough calls, it does not matter how flat the structure is. Progress will slow. Self-led teams that commit to learning together grow stronger. Leadership is not a role they wait for. It is a practice they share.

What about human–agent teams?
Here, the law shows up in different ways. A team that works with automated systems cannot treat the work as finished once the tool is in place. They need to keep learning. What works, what fails and what changes. They need to take responsibility not just for the outcome, but for how they got there.

These teams succeed when people lead the process. When they check assumptions. When they notice the early signs of friction. When they adapt. It is not about mastering the tool once. It is about improving the way humans and systems work together over time.

Lukas was a mentee of mine. He worked for a mid-sized manufacturer in the Black Forest. Thoughtful, methodical, deeply committed to doing things right. When his company decided to shift to agile, self-organised teams, Lukas took the lead. He spent weeks shaping a rollout plan that was clear, structured and easy to support. Leadership approved it without changes. The kickoff went also well, people nodded and the direction seemed set.

After another slow meeting, I asked him how he thought things were going. He hesitated, then admitted he was surprised. “I thought the plan would carry us further.” I said, “You don’t need another plan. You need to navigate.” He looked at me for a long moment, then nodded. That was the turning point. Lukas shifted his focus. He started checking in directly, listening more closely, noticing where people were stuck and why. He clarified expectations in smaller steps, helped the teams find early wins, and paid more attention to how the change felt, not just how it looked on paper. Slowly, things began to move again. In my experience is leadership not the plan. It is what happens after the plan meets reality. It is how you stay with people through resistance, confusion and hesitation; not by controlling the outcome, but by guiding the process.

In self-led teams, that kind of guidance still matters. When no one is steering alone, someone still needs to name what is unclear and what is missing. If no one holds the thread, even strong teams begin to drift.

In human–agent teams, navigation is less visible but just as important. A plan might include tools and workflows, but no system checks itself. Someone has to make sense of the results, notice when something no longer fits, and decide when to adjust. If that work is missing, things do not fail all at once. The team gradually loses connection to the goal.

One of the hardest moments in my leadership life involved someone I trusted deeply. A close friend. We had built the company together. Shared long nights, clear goals, and the feeling that we were building something that mattered.

Then came a decision we could not agree on. He wanted to move ahead with an idea I believed was too risky. We talked. Honestly, directly. But we stayed on opposite sides. In the end, the decision was mine. Not as a friend. As someone responsible for the company and the people in it. I said no. He was disappointed. Things between us changed. What had once been easy now needed more attention. For a while, I kept asking myself whether I had made the right call. Whether the cost was too high.

Looking back, I still carry that moment with me. But I no longer doubt what it taught me. Leadership means putting others first. Not by pleasing everyone, but by protecting what matters. Even when that means saying no to someone you care about.

Self-led teams need that kind of leadership too.
No one holds the title. But someone still needs to act.
Someone needs to speak up when something does not feel right and choose what serves the team instead of what feels comfortable. Someone needs to keep the long view when others look for shortcuts.

And in human–agent teams, the work remains human.
A system can help. It can take on tasks, but it cannot care, it cannot decide who is served and who is overlooked. It might act more fairly than a person would. But it can also reflect bias no one sees. That responsibility does not go away. It stays with us.

One question I come back to
Am I serving the team or protecting my own comfort?

Leadership is like walking on ice. As long as the surface holds, no one thinks about what lies beneath. People move freely, speak openly, challenge ideas, and take initiative. But the moment that surface feels thin, everything becomes tentative. Conversations lose depth, decisions take longer, and energy slips away. Not because people disagree with the goal, but because they no longer feel safe on the way there.

Trust is that surface. It holds more weight than plans, tools, or roles. When it is strong, teams can stretch, grow, and commit. When it weakens, the cracks are invisible at first, but they slow everything down. Jonas, a leader who worked for me, experienced this shift in his own team. He was respected, fast-moving (used to be an athlete), and full of ideas. On paper, everything looked right. His team delivered, meetings ran on time and feedback happened. But something didn’t sit right. The room was quiet. People nodded, but they rarely pushed back. Conversations stayed safe. One day, a colleague told him on a morning run, “I think people aren’t sure they can really be honest with you.”

That was the moment he realised what had been missing. He hadn’t broken trust through one big mistake. He had worn it down in small ways. By speaking too much and soon, by deciding without listening, by focusing more on the result than on the people working toward it. He used “I” when he should have said “we.” Not (only) out of ego, but out of habit.

He knew he needed to change. Not just to protect his role, but to rebuild what mattered. He slowed down. Asked more open questions. Paid attention when someone hesitated. Followed through when he said he would. Stayed in conversations that made him uncomfortable. Over time, the atmosphere in the team got „safer.“ People began to speak more directly and ownership grew. Not because they were asked to step up, but because the ground beneath them felt solid again.

In self-led teams, trust is not a luxury. It is what makes shared responsibility possible. Without it, people stay cautious, avoid conflict, and keep their best thinking to themselves. With it, they speak up, disagree well, and move together.

In human–agent teams, trust takes on a new dimension. Systems might handle tasks, but they do not take responsibility. People still need to know who defines the rules, who checks the output, and what happens when something breaks, which often does (especially with AI agents). The issue is not just technical. It is relational. Trust here means clarity, transparency, and a sense of agency. Without that, even the smartest system becomes a source of uncertainty.

Final thought
Leadership that lasts begins with trust.
It grows slowly, quietly, without applause.
But once it holds, it carries everything else.

Law Nr. 7 coming soon