AI Takes Over Tasks, Not Roles: Why AI Is an Organizational Question

At Business Forum 2026 I argued a thesis that goes beyond tools and efficiency: AI is also an organizational question. It changes not only how fast we work, but how work itself is organized. This article sums up why classic departments reach their limits, and how a multidimensional model answers that.

88%

of companies use AI in at least one business function

McKinsey Global AI Survey, 2025

90%

report no measurable productivity effect from AI

NBER, February 2026

21%

of German companies have an AI strategy

Bitkom, 2026

Contents

• A look back: the digital role model of 2019
• AI takes over tasks, not roles
• Three phenomena
• Not just an Implementation Gap: an organizational gap too
• From pyramid and matrix to multidimensionality
• Two examples: mechanical engineering and pharma
• Three questions for tomorrow morning
• Frequently asked questions
• Sources

A look back: the digital role model of 2019

In 2019 we developed a “digital role model” for a regional daily newspaper publisher. Weeks of preparation, dozens of conversations, and in the end a new org chart: a product manager for the website, a topic manager, a channel manager: roles that hadn’t existed in that form, together with processes and qualification programs.

Looking at that model today, AI can already take over a large share of the tasks we defined back then. Crucially, people from entirely different areas can perform those tasks. Someone from the digital unit now handles work in editorial or ad sales. What is tied to personal contact stays: field sales, consulting, reporting. Much else has become interchangeable between units.

AI takes over tasks, not roles

That is my first anchor. AI takes over not roles but tasks: researching, writing, designing, analyzing. And those tasks run across the org chart. AI doesn’t care about department boundaries, and neither do the employees using it: whoever can complete a task faster themselves won’t first take the detour through five departments.

Responsibility for decisions stays with people. But the executing tasks detach from the boxes in the org chart. That is what puts classic structures under pressure.

Three phenomena

We see three effects in almost every company, including our own:

• Skill inflation. A junior with good AI skills quickly reaches senior-level results. Classic career and salary paths no longer hold cleanly. Promotion by years was always questionable; AI makes it visible.
• Professions blur. Designers develop, marketers produce videos. Good people always crossed disciplinary lines. Now it becomes the norm, without the former friction.
• Loss of purpose for whole departments. When capabilities are no longer exclusive, the question arises: why this organizational unit at all? The professional grouping stays useful, but as the top ordering principle it no longer holds.

Not just an Implementation Gap: an organizational gap too

Despite broad usage, 90% of companies report no measurable productivity effect from AI (Source: NBER, February 2026). One reason is the Implementation Gap: the missing operating infrastructure. My addition: it is also an organizational gap. AI will act across org charts whether we want it to or not. The only question is whether we shape it or let it run under the hood.

AI doesn’t force us to rethink organization. It just makes visible what we should have done long ago.

Even once the tool question is solved via AI Operations, the org question remains. Whoever steers it keeps control while giving AI room to take effect.

From pyramid and matrix to multidimensionality

The classic pyramid has strengths: clear responsibility, fast decisions at the top, unambiguous escalation. But it is slow: in a dynamic environment, good communication from the bottom up arrives too late. Yet that is what we need with AI, because valuable things constantly emerge at the base.

The matrix is better suited, with more perspectives and more innovation, but has its known weaknesses: dual reporting lines, unclear responsibility, too many meetings.

Our organizational models are built for stable environments. AI shows how dynamic work has long become. Our answer is four dimensions:

1. Team: value creation. The leading element: people are organized where they create value, not by profession.
2. Discipline: craft. The former discipline stays, to develop skills and to judge AI output professionally.
3. Pillar: internal future-readiness. Load-bearing columns like brand marketing or administration, recruited from all teams.
4. Squad: change on a timer. Temporary, cross-disciplinary project units.

The biggest shift: the primary ordering principle is value creation (team), no longer craft (discipline). At netzstrategen the three teams are Consulting, Client Operations, and Platform & Technology; the disciplines Strategy, Search, Content, Development, and Design/UX each send people into every team. That such solutions emerge with the team rather than for it is no accident. See Built with the Team.

Two examples: mechanical engineering and pharma

What this looks like outside a consultancy is shown by two examples. Sven, online marketing manager at a machine builder with 250 employees: team “Plant & Product” (he sits with product development, good for go-to-market), discipline “Content & Campaigns”, pillar “Employer Brand”, squad “Website Relaunch”. One box on the org chart, four axes in daily work.

Sara, brand manager at a pharma company with 1,400 employees: team “Product & Brand”, discipline “Brand Marketing”, pillar “Compliance & Approval” (medical-legal-regulatory), squad “Product Launch”. In the larger group it takes more loops. The core principle stays: organize around value creation, not around profession.

Three questions for tomorrow morning

The first step is not a reorganization but visibility. Three questions help to start:

1. Which roles exist only on the org chart?
2. Which employees already work across several dimensions but have only one box?
3. Which “projects” are in truth permanent fixtures that belong in the regular organization?

The second big change is how AI changes outward visibility. How the organizational shift connects to it is shown in the recap of both Business Forum talks. That tools achieve little when the work logic stays the same is explored in Workflow first, tool second.

Frequently asked questions

Does AI replace whole roles and departments?

No. AI takes over individual tasks, not whole roles. Decisions stay with people. But because tasks run across the org chart and can be performed by many, organizing by functional department loses its ordering power.

What is the organizational gap?

In addition to the Implementation Gap (missing operating infrastructure), the organizational gap describes how AI changes structures as long as companies keep steering by craft instead of value creation. Whoever shapes this actively keeps control.

Isn’t this just a double matrix?

Partly, but with clear guidance: the value-creation team is the leading element, the discipline serves craft development. The team holds capacity; outcome comes via squads. Where the biggest lever sits is shown in a free introductory call.

Sources

• NBER: Firm Data on AI, February 2026
• McKinsey: The State of AI, Global AI Survey, 2025
• Bitkom: Artificial Intelligence in Germany, 2026