AI-enhanced managers: transforming teams

Managers augmented by AI They use generative tools to better prepare decisions, organize work and support teams, without delegating their responsibility or the quality of human relationships.

Artificial intelligence has permeated organizational life to such an extent that it has silently transformed the dynamics of cooperation, the very nature of work, and the role of those who manage it. Initially perceived as a technical advantage, it is now reshaping the relational grammar: it is no longer just the tasks that are changing, but also the way we communicate, make decisions, and work together. As the algorithm becomes an invisible intermediary—summarizing, reformulating, and prioritizing—communication becomes less functionally burdensome while sometimes losing its human touch. The collective becomes more fluid, but risks becoming more fragile: what circulates is no longer the raw voice of individuals, but a condensed version of filtered signals. Faced with this subtle dematerialization of the link, the manager discovers himself as the guardian of relational density: he identifies silences, restores dialogue outside the machine, invents rituals where meaning and affect are reintroduced into a world that has become too clean.

In so-called "augmented" teams, this algorithmic mediation leads to a paradox: less effort to coordinate, but a greater risk of disengagement. When reports are automatically generated and schedules are calculated on their own, the temptation is strong to minimize human presence; yet, creativity, learning, and trust often arise from friction, disagreements, and informal detours. Some organizations have recognized this by deliberately reintroducing "AI-free" spaces: meetings without summaries, debates without scripts, and listening time where words aren't immediately polite. These breathing spaces maintain the vitality of the group by allowing room for fruitful misunderstandings and unfinished ideas that spark innovation.

This reconfiguration of communication is just a taste of a deeper transformation: the operational autonomy provided by AI shifts the center of gravity of management. When a consultant, analyst, or engineer can complete most of their workflow with the assistance of a predictive engine, the manager's role shifts from controlling resources to regulating their use. It's no longer about verifying the volume of data processed, but about assessing the relevance with which the tool is used, the ability to question its suggestions, and the capacity to add human value. Traditional indicators—raw productivity, turnaround time—become insufficient; it's now necessary to measure discernment, the quality of reasoning, and the ability to accept the degree of uncertainty that the machine cannot eliminate.

This transformation is accompanied by a more subtle phenomenon: the intensification of cognitive load. By delegating repetitive tasks, AI leaves the most complex segments to employees; the mental flow becomes denser, more continuous, and sometimes exhausting. Recent studies in organizational neuroscience show a decrease in brain activity during the uncritical use of generative assistants, followed by poorer information retention and a diminished sense of ownership. In other words, the autonomy produced by automation can turn into imposed autonomy: one is left alone to face complexity, lacking the reference points to regulate effort or slow down. Managers must therefore learn to detect cognitive fatigue, establish room for maneuver, and value periods of slow reflection where the pace imposed by the machine is suspended.

To this equation is now added a new player: the AI ​​agent, an algorithmic companion capable of acting, communicating, and sometimes initiating decisions. Neither a simple tool nor a true colleague, it blurs the lines between human and technical resources. Its presence raises questions about responsibility: who is accountable for an action triggered by an autonomous system? It challenges trust: how can teams be convinced that this agent is an ally and not a rival? It disrupts coordination: should its work be monitored, corrected, or simply interpreted? Pioneering organizations are already seeing the emergence of new roles – “human/AI mediator,” “prompt curator,” “coexistence orchestrator” – tasked with optimizing this interaction without undermining human collaboration.

But wherever AI makes its presence felt, the temptation to ban it arises in response. Regulated sectors—finance, pharmaceuticals, public institutions—have preferred to freeze its use rather than confront the risks of data leaks, non-compliance, or a decline in work quality from the outset. These moratoria do not reflect a rejection of the technology; they primarily reveal the need for a robust governance framework, where security, ethics, and training precede deployment. In the long run, bans will likely remain temporary: the challenge is to shift from a defensive approach to one of responsible adoption, based on supporting managers and clarifying legitimate use cases.

From a cultural perspective, the major risk is not substitution but homogenization: if everyone relies on the same generative models and settings, the diversity of styles, ideas, and approaches could be flattened. To preserve the cognitive richness of an organization, it becomes essential to promote personal writing over assistance, divergence over convergence, and to encourage everyone to cultivate their uniqueness instead of conforming to the standardized prose of the algorithm. Management has a key role here: recognizing originality, valuing creative risk-taking, and rewarding nuance rather than mere efficiency.

These changes call for an ambitious skills development program: technical mastery to understand the limitations and biases of existing models; socio-emotional intelligence to maintain the quality of relationships; and an ethical framework to navigate dilemmas of responsibility and transparency. The goal is less about training battalions of data science experts than about shaping "trust architects," capable of orchestrating a hybrid ecosystem where different forms of intelligence complement each other without competing. In this context, the manager becomes a process designer, a coach of discernment, a guarantor of meaning, and a regulator of mental effort.

For organizations with a strong social mission, such as those in the public or healthcare sectors, the stakes are even higher. They must ensure that the promise of performance does not overshadow the quality of relationships, the fairness of decisions, and the preservation of public trust. This requires structured experimentation, systematic documentation of effects, and the establishment of forums where feedback from the field informs strategy. The deployment of AI cannot be driven solely by the pursuit of quick wins; it must be integrated into a process of continuous improvement, fueled by critical observation and team participation.

Ultimately, artificial intelligence does not spell the end or simplification of management; it increases its strategic scope, complexity, and profoundly human dimension. Designing relational rituals in the age of automation, evaluating work when the data is shared between humans and machines, protecting the cognitive integrity of employees while leveraging algorithmic power: these are all challenges that demand renewed leadership. More than ever, value will reside not only in technical mastery, but in the ability to orchestrate a space of trust where humans remain a source of initiative, interpretation, and meaning, even when the machine performs the bulk of the apparent work. The future of work will only be sustainable if algorithms remain a partner, and never the silent conductor that silences the plurality of voices.

Implementing AI-enhanced management

AI-enhanced management relies less on automating decisions and more on better preparing them. Managers can use generative AI to synthesize information, explore multiple scenarios, prepare for meetings, or identify early warning signs in a project. Responsibility, arbitration, and team relationships remain human.

The deployment must combine rules of use, verification of results, data protection and collective learning.

To learn more: see our analysis of workslop and our article on AI in the service of sustainable work.