The New Division of Labor Between Humans and Machines

Economists Ajay Agrawal, Joshua Gans, and Avi Goldfarb describe artificial intelligence as a technology that dramatically improves one capability above all: prediction.

Algorithms can now calculate with impressive accuracy:
which products customers are likely to buy
how prices affect demand
the risk of loan default
which supply chain disruptions are to be expected.

However, strategic decisions consist of more than forecasts. They always involve judgment—that is, the assessment of conflicting objectives, risks, and long-term consequences.

AI can calculate how likely an event is.

But it cannot decide:
which risk is acceptable
which strategy makes sense in the long term
which priorities a company should set.

These decisions remain a management responsibility.

Uncertainty as the new normal

Strategic decisions are rarely made under stable conditions today. Markets are changing faster than ever before due to technology, regulation, and geopolitical developments.

Companies must make decisions even though key factors remain unclear:

• New technologies are changing industry structures
  Political decisions are influencing markets
  Customer needs are evolving dynamically
  Global supply chains remain fragile.

This situation is often described by the term VUCA – Volatility, Uncertainty, Complexity, and Ambiguity.

In such environments, a traditional decision-making model loses its effectiveness: analyze first, then decide. Instead, organizations must decide and learn simultaneously.

Strategy thus becomes less of a long-term master plan and more of a process of continuous adaptation.

The Limits of Automated Decisions

The idea of ​​fully automated business decisions underestimates a key characteristic of modern AI systems: their inconsistency.

Recent research describes the phenomenon of so-called “artificial jagged intelligence.” AI systems can demonstrate outstanding performance in certain areas, but react surprisingly unreliably to seemingly similar problems.

For organizations, this means that AI remains an extremely powerful tool—but not an autonomous decision-maker.

Especially in situations with high consequences—such as strategic investments, regulatory risks, or crises—the value of human assessment increases.

The manager of the future will not replace AI.

They will calibrate its results, evaluate its limitations, and translate analyses into responsible decisions.

The Three Dimensions of Irreplaceable Leadership

Research on AI, decision-making, and organizational transformation reveals three core competencies that form the foundation of modern leadership. Together, they explain why leaders remain indispensable even in the age of intelligent machines.

1. Judgment under Uncertainty

Artificial intelligence can calculate probabilities – but strategic decisions are rarely based on probabilities alone. They involve conflicting goals, values, risks, and long-term consequences.

Leaders must make decisions even when key information is missing or contradictory.

This ability is based on three factors:

Experience, which makes patterns recognizable over many years
Contextual understanding, which classifies economic, political, and cultural factors
Intuition, i.e., condensed experiential knowledge

The crucial question is therefore not:

What does the model say?

But rather:

What consequences are we prepared to bear?

This form of judgment remains a profoundly human competence.

2. Providing Orientation
Uncertainty creates not only analytical but also psychological challenges.

Employees expect guidance from leaders: a plausible explanation of where an organization is headed and why certain decisions are made.

Leadership therefore also means making complexity understandable.

A good leader
explains assumptions and risks
makes conflicting goals transparent
articulates a coherent direction.

Especially in times of technological transformation—for example, through artificial intelligence—this ability becomes a key source of organizational stability.

3. Mobilizing People
Perhaps the most important leadership skill is generating collective action.

Organizations don't automatically follow the analytically best solution. They follow decisions that are credibly communicated and supported by leadership.

Strategies therefore rarely fail due to flawed analysis—but rather because people don't support them.

An effective leader thus creates three prerequisites:

Trust
People must be able to trust that decisions are made responsibly.

Participation
Teams must understand their role in implementation.

Purpose
Employees are more likely to follow strategies when they understand why a decision is important.

This social dimension of leadership cannot be automated. It is based on legitimacy, communication, and personal credibility.

Practical Examples: Leadership in the AI ​​Age

Many successful companies are already demonstrating what this new form of leadership looks like.

Amazon: Decisions as Hypotheses
Amazon operates on the principle of "two-way door decisions." Many decisions are deliberately made quickly and then tested.

The leader doesn't create perfect planning, but rather rapid learning cycles.

Microsoft: Culture as a Leadership Task
Under CEO Satya Nadella, Microsoft consciously changed its corporate culture—moving away from internal competition toward a learning-oriented approach and collaboration.

The decisive lever of the transformation was not technology alone, but leadership and culture.

Nvidia: Strategic Judgment
Nvidia invested heavily in AI chips early on—long before the market for generative AI exploded.

This decision was not based on certain forecasts, but on strategic judgment under uncertainty.

Leadership as a Competitive Advantage

The more data-driven companies become, the clearer a paradoxical truth becomes: The limiting factor in decision-making is rarely the analysis itself, but rather its implementation.

Many organizations today possess excellent data, powerful models, and sophisticated analytics. Yet, transformations frequently fail.

The reason rarely lies in the technology. It lies in leadership.

International labor market studies therefore show that competencies such as analytical thinking, resilience, systems thinking, creativity, as well as leadership and social influence are among the most important skills of the future.

Technology does not replace leadership—it makes its importance more apparent.

Leadership Takeaway

Perhaps the greatest irony of the AI ​​revolution is this:

The more powerful machines become at analysis, the more crucial the human contribution to decision-making becomes.

Artificial intelligence can generate forecasts, calculate scenarios, and simulate options. But it cannot determine which risks are acceptable, which goals take priority, or why an organization should pursue a particular path.

Therefore, the manager of the future is not the human being who calculates against machines.

It is the human being who leads where calculation alone is insufficient: in situations of uncertainty, conflicting objectives, and collective decision-making.

This is precisely what makes leaders indispensable, even in the age of artificial intelligence.

Quellen und Studien

Agrawal, A., Gans, J., Goldfarb, A. (2018).
Prediction Machines: The Simple Economics of Artificial Intelligence. Harvard Business Review Press.

Agrawal, A., Gans, J., Goldfarb, A. (2018).
Prediction, Judgment and Decision-Making. National Bureau of Economic Research (NBER Working Paper 24243).

Brynjolfsson, E., Li, D., Raymond, L. (2025).
Artificial Jagged Intelligence: AI’s uneven capabilities. National Bureau of Economic Research (NBER Working Paper).

McKinsey Global Institute (2025).
Superagency in the Workplace: Empowering People to Unlock AI’s Full Potential.

Stanford University – Institute for Human-Centered AI (2025).
AI Index Report 2025.

World Economic Forum (2025).
Future of Jobs Report 2025.

Harvard Business Review – verschiedene Beiträge zu Decision-Making und Leadership im KI-Zeitalter.