For decades, every major technological innovation has been accompanied by an implicit promise: to simplify the world. Automate, accelerate, optimize — verbs that suggest a reduction of complexity and greater clarity in action.
Artificial intelligence seems to embody this promise at an unprecedented level. If it allows us to automate an increasing number of tasks, shouldn’t it make our decisions simpler, more rational, more efficient?
In reality, the opposite is probably happening.
There is a widespread belief that automation simplifies. Yet while AI dramatically reduces execution costs, it does not reduce decision costs, it increases them.
When acting becomes inexpensive, the world suddenly fills with a profusion of possible actions.
Before AI, launching a new product could require months of preparation. Simulating a marketing campaign involved significant time and financial investment. Testing multiple variations was a luxury reserved for elite organizations.
Today, an AI system can generate and compare thousands of variants in real time, with immediate probabilistic feedback.
Companies no longer face a single market, but a multitude of potential micro-markets, millions of possible strategic configurations, and therefore just as many decisions to make.
This is what I call hypercomplexity.
hypercomplexity
Hypercomplexity does not simply mean more data or more options. It refers to a level at which it becomes impossible for a human being to simultaneously grasp all possibilities. Decisions can no longer be labeled simply as right or wrong; decisions become trajectories evolving within a dynamic, continuously shifting space.
Where once one chose between A or B, AI now generates strategic sets such as {A1, A2, A3… A1000}, constantly recomposing as competitors use similar capabilities.
Here a paradox emerges: the more numerous and accurate AI predictions become, the more unpredictable the overall system grows.
When all companies rely on comparable AI models, their decisions influence one another. Feedback loops arise, strategies mirror each other, and adaptation must occur in real time.
Rationalized decision-making does not stabilize markets, it increases their dynamism and instability.
In complex systems, hyper-adaptation generates turbulence. And when changes occur too rapidly, the system no longer has time to return to equilibrium.
This transformation will likely have two major consequences for business management.
Consequences for business management
First, the emergence of a different managerial profile: less focused on isolated components (whether technologies, methods, or processes) and more systemic, capable of orchestrating interactions and building adaptive human decision teams.
Second, a profound shift in management models. Protocols are designed for relatively stable and repetitive environments. They work when one can say, “if X, then Y.”
But what happens when X never occurs in exactly the same way, when X may take a hundred different forms, and when the time available to decide approaches zero?
In such a context, protocols can become prisons — rigid frameworks within which organizations eventually suffocate.
Conclusion
Viewing AI as merely a tool that will allow us to do the same things faster and better is therefore an illusion. AI is not a new steam engine that we can integrate without transforming ourselves.
AI signals the entrance into a new world — a world to which it will not be AI that must adapt, but us.
