Lately, much of the conversation has focused on what machines can do better than humans: analyze data, write reports, develop strategies, and support decision-making. And when the discussion turns to what sets humans apart from machines, it often feels as though we are auditioning for a panel we appointed ourselves. The same answers tend to emerge: creativity, empathy, critical thinking, or judgment. That says Sonsoles Alonso, a consultant, systems thinker, contemporary musician, and founder of Composition8.
But there is something we forget to mention: perception. You walk into a meeting room, and you already know. The atmosphere is different. People’s shoulders carry a burden that wasn’t there yesterday, and the eyes avoiding each other say more than an hour of meetings ever could.
No one has spoken a word, yet you feel it in your throat or in that place where things land before we have language for them. AI cannot do that, because AI waits for data. And you are the data.
AI operates in the world of language and patterns. It can describe what is happening. It can read thousands of books on trust, collaboration, and leadership and tell compelling stories about conflict and culture. But it cannot feel what exists between people. In addition to language, humans possess a unique source of information: the body.
AI does not have that, and our bodies are often the first to know. Long before we find the words, our bodies comment on what is happening: through tension, through the way our shoulders tighten, through the feeling that a conversation is not flowing as smoothly as usual. People disengage before they actually leave, and tension can be felt before anyone expresses it. A leader who picks up on these signals has access to information that no AI dashboard can provide.
The same applies outside the meeting room. A geopolitical event does not affect a market only when the numbers confirm it. The first signals often appear much earlier: in conversations with customers, in uncertainty among suppliers, or in questions that suddenly come up more often.
If I were to summarize the differences between humans and AI in broad terms, the table would look something like this:
| AI | Human | |
| Primary capability | Semantic processing | Semantic + somatic processing |
| Understands language? | Yes | Yes |
| Has a body? | No | Yes |
| Feels tension, fear, trust, or uncertainty? | No | Yes |
| Participates in power dynamics? | No | Yes |
| Creates meaning? | Statistically | Embodied and social |
| Acts under uncertainty? | Simulates | Experiences |
AI is a semantic machine: it predicts relationships between symbols. Humans, by contrast, are both semantic and somatic. We understand the world not only through words and concepts but also through what we feel, perceive, and experience. Semantic information is the world of meaning: words, concepts, explanations, stories, strategies, and models. When ChatGPT talks about leadership, it operates entirely within that domain. It can describe trust, but it cannot experience it; it can explain conflict, but it cannot stand in the middle of one. AI works with meaning; leaders create meaning.
Somatic intelligence concerns what we feel before we can explain it: gut feelings, emotional resonance, and bodily responses. When a CEO walks into a boardroom and immediately senses that something is wrong before anyone has spoken, that is somatic intelligence at work. The information first appears as an experience and only later becomes language.
In a stable world, you can rely on plans, models, and forecasts. But a complex world cannot be predicted. Leaders must navigate geopolitical shifts, economic uncertainty, and the impact of AI. New patterns do not appear in advance within a plan; they emerge from interactions between people and events in real time. In systems theory, we call this emergence. The whole reveals something that cannot be seen in any of the individual parts. This makes perception a practical necessity.
The leaders of the future will not distinguish themselves through better answers. They will distinguish themselves by perceiving signals before others do.
Leadership as avant-garde
Avant-garde is often mistaken for doing things differently for the sake of being different. But the word comes from French and originally referred to the vanguard of an army: the soldiers who moved ahead to explore unknown territory. Their task was not to be different, but to see what was coming before anyone else.
That is what makes the concept of avant-garde relevant to leadership. As a concert pianist in contemporary music—the avant-garde of the musical world—I have experienced this meaning firsthand. I have gone to places others had never been before.
As a leader, you move through terrain that has not yet been fully mapped. In a complex world, the past becomes an increasingly unreliable guide to the future.
This is also where AI reaches its limits. AI is impressive at recognizing patterns from the past because it is trained on existing data. AI primarily knows what we already know. But unknown territory cannot be reduced to historical patterns. When contexts change, new technologies emerge, or markets behave unexpectedly, there is no dataset that tells us what to do. That is when human leadership becomes essential, because leaders can perceive signals, create meaning from uncertainty, and choose a direction when no map yet exists.
In that sense, leadership today is a form of avant-garde: not following what is already visible, but perceiving what is beginning to take shape and acting on it.
The leader as a composer of complexity
Many people think music consists of melody, harmony, and beautiful instrumental sounds. But for composer Edgard Varèse, music begins where that definition ends. For him, sirens, machine noises, electronic sounds, and noise itself are also musical material. With his famous statement, “music is organized sound,” he makes clear that music is not about organizing notes, but about organizing sound.
Today’s market resembles Varèse’s world. There is no recognizable melody, only a continuous interplay of contradictions, geopolitical shifts, and technological changes interacting before we fully understand them.
The more interests, dependencies, uncertainties, and human dynamics interact, the harder it becomes to capture reality in an AI prompt. The leader’s task is not to eliminate complexity but to organize it: to determine which signals matter, which patterns are meaningful, and which questions should actually be asked. Only then does AI become useful. This requires something different from leaders than analysis alone. In a market full of noise, you are not a conductor of harmony; you are a Varèse composing meaning from complexity.
When I enter organizations, I often encounter a tangle of interests, concerns, and perspectives. Everyone hears something, but not yet the same music. In the collaboration between Composition8 and MDFT Pro, this is the starting point of the two-day AI Fixer Sprint. On the first day, we bring teams and stakeholders together and use systemic process tools to create a shared understanding of what is happening. What problem are we solving? For whom? Which interests are involved? How does power manifest itself? Who or what influences whom?
Only once these questions are clear does it become visible which AI solution truly adds value. The work, therefore, is not about eliminating noise. Just as Varèse organized sound into music, today’s leader organizes complexity into direction. Collaboration is organized complexity.
From meaning to meaning-making
AI works with meaning; leaders create meaning. This distinction becomes more important as the world grows more complex. AI can recognize patterns, suggest options, and process enormous amounts of information. But it does not live within the reality that information describes. It does not participate in relationships, feel tension within a team, or experience uncertainty.
Leaders do. The highest form of leadership today is creating space for people to express what they see, making tension discussable, and recognizing patterns before they become problems. Ultimately, AI can help explore possibilities, but it does not choose a direction. That choice remains human work.
This is your leadership challenge in the age of AI: not to dictate the truth, but to see emerging signals earlier, create meaning together, and choose a direction.
