A company without managers: three traps companies fall into when implementing AI

From 50 managers at ProIT Fest, only three raised their hands: decision-making with AI became easier. The number of managers in companies is shrinking — but the work isn't getting simpler. This is the first part of a series about how AI is changing the nature of management and why most companies are still moving in the wrong direction.

Three Traps of AI Management

Most companies approach AI transformation with a single expectation: eliminate unnecessary layers, accelerate decisions, reduce the management apparatus. The first results look convincing — headcount shrinks, tools work, reports write themselves. But then something strange begins: more decisions, less clarity.

Automation trap. AI takes on routine work, but doesn't take on responsibility. Managers don't disappear — they shift into a mode of overseeing algorithms and add a new layer of approvals on top of the old one. The process accelerates in form, but not in substance. In the end, there are fewer managers, but no less bureaucracy.

Speed trap. When a decision can be made in a minute, there's a temptation to do exactly that. Context is lost, nuances are ignored. A company gets many quick decisions — and some of them turn out to be wrong, because no one stopped to think about the consequences.

Power vacuum trap. When there are fewer managers, formal hierarchy breaks down faster than a new one can form. Informal influence concentrates among those who mastered the tools best — not among those who think most strategically. Hidden power centers emerge that are difficult to see and even harder to correct.

Framework of Three Dimensions

The authors propose evaluating management transformation not by a single metric, but by three parameters at once. Most companies excel at the first — and lose on the other two.

• Speed of decision-making — AI really does help here. This dimension grows in almost every company that implements automation, and it's often the main argument in favor of AI.

• Quality of decisions — the picture is mixed. Analytical tools provide data and reveal patterns, but create a false sense of certainty where managerial intuition and context — which AI lacks — are needed.

• Distribution of responsibility — the most vulnerable dimension. When AI recommends and a human approves, responsibility becomes blurred. No one feels like the true author of the decision — and this creates a serious cultural risk.

It's this third dimension that explains the ProIT Fest statistics. Technically, many companies have AI tools. The feeling of actually being in control — only a few. The most honest question for a team: when was the last time you made a difficult decision where AI didn't provide a ready-made answer?

Where Your Company Stands

At the end of the first part — a practical test that evaluates not the level of technological equipment, but the management maturity of the organization. Three key questions:

• Who is responsible when an AI recommendation turns out to be wrong?
• Do you have managers who can disagree with an algorithm — and can justify it?
• What decisions in your company can AI fundamentally not make?

Honest answers to these questions are usually uncomfortable — but they're exactly what show where the real gap lies.

"There are fewer managers — but responsibility hasn't gone anywhere.

It's just been redistributed chaotically," — the key thesis of the article.

What This Means

A company without managers — is not utopia or absurdity. It's a realistic prospect for some functions and processes. But the path there doesn't lie through cutting people, but through rethinking what exactly a manager does: does he make decisions — or just service a system that makes them?

For now, most organizations are in the stage of illusions: there are fewer people, but organizational complexity hasn't decreased.