When AI handles planning, coordination and evaluation, what remains for the leader? This edition explores what is, by definition, absolutely non-delegable.
Prefer watching to reading? Watch the full video — then come back here for the detailed references.
What happens to leadership when its traditional functions are taken over by systems that never sleep, have no ego and do not make the typical mistakes of someone under pressure? This is not a philosophical question about a distant future. It is a tension that is already quietly settling into organisations in 2026.
01 — ContextFor decades, organisational theory defined leadership through measurable behaviours: setting direction, aligning teams, making decisions under uncertainty, giving and receiving feedback, identifying talent, managing conflict. Each of these functions was, at some point, considered exclusively human.
The problem is that each of them is now being partially replaced or assisted by AI systems. Workforce analytics platforms anticipate employee attrition risk before any leader notices it. Performance analysis tools generate evaluations based on digital behavioural data. Project management systems identify bottlenecks, redistribute resources and suggest priorities. Language models draft internal communications, summarise meetings and propose action plans.
If leadership is simply the sum of these tasks, then the question is legitimate: what is the leader for?
The hasty answer would be to say that the leader becomes an "AI manager", someone who oversees the systems, validates outputs and ensures process quality. This narrative is comfortable but dangerous, because it reduces leadership to a control function and ignores the dimension that no system can replicate: presence.
A leader's presence is not the sum of their decisions. It is the capacity to create context, to give meaning to what is happening, to transform uncertainty into a narrative that the team can understand.
"The differentiator in high-performing teams isn't the sophistication of their tools, but the coherence between a leader's words and their actions."
AI can suggest what to do. It cannot explain why it is worth doing.
03 — ParadoxThere is a dominant narrative in the corporate world today that describes AI as a way of "augmenting" human capability. The AI-enabled leader would be faster, better informed and more effective. But this argument conceals a real tension.
When a leader receives an algorithmic recommendation and follows it, who is actually leading? When a performance management system generates a rating and the leader communicates it without questioning it, who is accountable for the decision? When a recruitment tool filters candidates and the leader only interviews those approved by the model, who defined the criteria that determine talent?
Gallup's 2025 reporting shows that only 23% of workers globally are actively engaged at work. The most relevant finding is not the absolute figure, but the fact that the human relationship with the direct leader remains the factor with the greatest explanatory power for the difference between committed teams and passive ones.
No AI system replaces this relationship. But many AI systems are replacing the moments in which this relationship could be built.
One of the least discussed effects of organisational automation is the compression of the informal spaces where leadership actually takes place. The conversation in the corridor before a meeting. The spontaneous team lunch. The moment when a colleague shares a concern that would not fit into a structured feedback form.
From a process efficiency standpoint, these moments are seen as noise. From the standpoint of effective leadership, they are the most valuable signal available.
Researcher Ethan Mollick at Wharton School has argued that AI is revealing a fundamental divide within organisations between those who use these tools to think better and those who use them to think less. The leader who uses AI to free up time and space for the conversations that systems cannot have is genuinely augmenting their own capacity. The leader who uses AI to avoid those conversations is hollowing out their own role.
05 — LiteracyThere is a dimension of this transformation that organisations have not yet addressed systematically: the need to develop a new form of leadership literacy specifically oriented towards decision-making contexts augmented by AI.
This is not about knowing how to use the tools. It is about knowing when not to use them. About understanding which types of decisions lose quality when delegated to a system. About recognising the moments when human presence is irreplaceable, not for sentimental reasons, but for functional and ethical ones.
The World Economic Forum's Future of Jobs 2025 report estimates that leadership and social influence skills will be among the most valued by 2030, precisely because automation will make the capacity to coordinate complex human contexts increasingly scarce. Yet the same report warns that leadership development programmes continue to focus on technical and managerial competencies, ignoring the relational and contextual dimension that distinguishes a process manager from a genuine leader.
06 — Non-delegableThere is a set of responsibilities that structurally belong to human leadership and that no technological advancement will transfer to a model. Moral accountability for decisions that affect people. The capacity to recognise distress before it becomes a data point. The ability to hold a team together in a situation of genuine ambiguity, where the model has no answer because the answer does not yet exist.
"Trust cannot be delegated. It cannot be optimised by an algorithm. It does not emerge from an engagement dashboard. It is built through consistency between intention and action, over time, in moments that are often not recorded by any system."
The real risk is not that AI will replace leaders. The risk is that leaders will allow themselves to be replaced by a version of themselves that delegates more and more and is less and less present, convinced that efficiency is a form of leadership.
"What is it that I, and only I, can give my team that no system will ever provide?"
This remains, inevitably, a question about humans before it is a question about machines.
Only 23% of employees worldwide are actively engaged at work — and the direct manager relationship is the most decisive factor.
Automation will reshape 170 million roles by 2030, making leadership and social influence the scarcest and most valuable skills.
AI adoption nearly doubled in one year, while AI-related incidents hit a record high — a sign that deployment is outpacing governance.
Reflections on technology, organisations and the people navigating both. Monthly editions.
Your data is protected. You can unsubscribe at any time.