Why AI demands more human processes
For some time now, Artificial Intelligence (AI) has ceased to be a futuristic promise for most companies. Today, it acts as one more engine of transformation. However, amid the fascination with algorithms, there is still a very current dilemma: how can this digital power be used without sacrificing the best of human connections, such as empathy and trust? To understand this problem and answer this question, we must first understand what AI should be and what it cannot be.
This is critical because, in general, the role of the traditional manager, still predominant in many companies, is overloaded with bureaucratic, redundant activities that have no clear connection with value creation. This work persists not out of real necessity, but because of organizational inertia, a lack of visibility into processes and the difficulty of distinguishing, in day-to-day work, what actually contributes to results. The effect is predictable: a significant portion of management’s time and energy is consumed without a proportional return.
It is estimated that between 30% and 50% of a traditional manager’s activities could be eliminated or replaced by AI, without much complexity, in the coming years. From a lean perspective, this estimate is consistent with the nature of today’s managerial work: a significant portion of time is still consumed by consolidating information, manual follow-ups, chasing people for deliverables and status meetings. These are, in essence, transactional activities, with low decision-making content and high repetition, and therefore largely open to digitization.
The crucial point, however, is not the percentage of tasks eliminated, but what must remain: the ability to prioritize problems, deepen the understanding of them, develop people and bring clarity to purpose. It is in this field that AI does not replace leadership. Its role is a different one: to help reveal the limitations of models still stuck in repetitive routines and to open space for a more demanding role, focused on building learning systems. In this sense, AI does not replace the leader, but it raises the bar by making the weaknesses in how they lead more visible.
This change, however, runs up against a still-dominant mental model, based on command and control. It is this frame of reference that defines how work is organized and how technology is incorporated into everyday life. When this logic prevails, AI tends to be used as an additional layer of efficiency, but applied on top of structures that remain unchanged. The result is predictable: speed is gained, but the way management operates is not transformed.
The differentiator is no longer in execution and shifts to the quality of one’s perspective: in how problems are defined, in the questions that guide the analysis and in the criteria used to interpret the answers. It is on this ground that AI meets lean leadership. Without this shift, the manager continues to operate from the same frames of reference, even with new tools at hand. The technology advances, but the logic of how they operate remains unchanged.
In this context, the time “freed up” by AI tends to be quickly taken up again by new demands and additional layers of coordination. This movement is less technological and more organizational, because in the absence of clear criteria about what should or should not be done, the available space is “naturally” filled.
To break this pattern, it is necessary to redesign the manager’s role. The gain is not in the time itself, but in the ability to make more discerning choices about where to intervene, what to eliminate and what to develop. Without this change, the additional capacity generated by technology dissipates, with no significant impact on the quality of management.
To do so, we must recognize that the line between technological support and human detachment is thin and often imperceptible. The use of AI becomes harmful when it begins to replace the direct connection between leaders and teams. By relying exclusively on “dashboards”, the manager moves away from the gemba. The software interface gains ground, while conversation, active listening and direct observation lose their place. The precision of data cannot replace presence. Without context or connection to the reality of the operation, numbers tend to lead to inevitably mistaken decisions.
The most critical effect of this process is not immediate, but cumulative. As direct interaction is reduced, teams begin to work with fragmented interpretations of reality. Problems are no longer understood in depth and come to be dealt with superficially. Over time, the collective ability to learn and improve the work weakens. In this environment, decisions rely increasingly on “indirect” signals, opening the door to interventions that seem precise but lack practical grounding.
As a result, a recurring sense of misalignment arises, with clear signs of this mismatch. Information becomes widely available, but loses its connection with the more complex meanings it carries. Feedback still exists, but it tends to become superficial, generic and excessively standardized, in the worst sense of the term, without the nuance necessary for human development. This environment weakens trust and limits collective learning, creating a much colder and more distant dynamic.
This “coldness” increases when precision is confused with understanding. AI is efficient at measuring and correlating, but it does not, by itself, access contexts, purposes or the complexity of the effort involved. The manager’s critical mistake is to treat metrics as absolute truths or as final verdicts. They should be starting points. For what? To broaden dialogue, deepen interaction, and guide investigations and trips to the gemba. Metrics should open conversations, not close them. When a number replaces dialogue, the organization loses contact with people and weakens its ability to learn from real work.
This kind of distortion exposes an important limit to the use of technology. Organizational culture is not formed from tools; it emerges from everyday choices, from interactions and from the way work is carried out. AI can capture signals and patterns, but it does not replace the shared experience that sustains these behaviors. Assessing this integration therefore requires much more than isolated indicators. It is necessary to observe whether the use of technology is strengthening shared understanding, the quality of interactions and the ability to learn from real work.
This balance reveals itself in observable signs in day-to-day work. Is it safe to expose problems and mistakes? Are interactions open and oriented toward understanding, or are they limited to transactional and defensive exchanges? Is leadership present at the gemba or does it operate from a distance? And are the tools serving the elimination of waste, better problem solving and value creation, or have they become an end in themselves? The answers to these questions indicate whether AI is improving the quality of the environment or reinforcing barriers between people.
In companies, examples are piling up of misguided uses of technology that end up deepening precisely the problems they were meant to solve. At the same time, more consistent experiences are beginning to emerge, showing that it is possible to integrate AI and management in a more coherent way.
In management, technology is not limited to simple adoption; it requires a review of the way work is structured. The manager’s role shifts toward organizing work so that technology supports the analysis and people lead problem solving and decision-making. By reducing the load of automatable activities, space opens up for what cannot be delegated.
Success will not lie in mastering algorithms, but in knowing how to use them to be more present, with more empathy, trust and focus on what really matters: people. AI amplifies the existing management logic. If management is distant, it will amplify that distance. If it is human and learning-oriented, it will act as a catalyst. The choice is not technology’s, but rests on what we want for our people.
