The Flattening: Why Telenor, Citi, and Microsoft Are Eliminating Management Layers

Jane Fraser at Citi cut management layers from 13 to 8. Telenor announced a "country-centric" restructure that removes regional management layers entirely. Meta, Intuit, Microsoft, and others have made parallel moves.

The pattern is consistent: management layers are being eliminated at scale.

This is not cost-cutting. It is structural redesign.

Why Layers Existed

Management layers served a purpose. In organizations too large for any single person to oversee, layers provided:

- Information aggregation: Work rolled up through layers, giving senior leaders visibility into what was happening below - Decision distribution: Authority was distributed across layers, so decisions could be made at appropriate levels - Quality control: Each layer reviewed the work of the layer below, catching errors before they reached the top - Career progression: Layers provided rungs on the ladder, giving workers paths to advance

For decades, this structure enabled scale. Organizations could grow to hundreds of thousands of employees because the layers managed complexity.

Why Layers Are Now Liabilities

The same layers that enabled scale now slow it down:

- Decision latency: Every layer adds time. In markets where speed matters, layers create bottlenecks that competitors without them do not have. - Information distortion: As information passes through layers, it gets summarized, filtered, and distorted. Senior leaders see a version of reality, not reality itself. - Accountability diffusion: When decisions pass through multiple layers, accountability becomes unclear. Everyone approved. No one owned. - Talent frustration: High performers chafe at layers that slow their work and limit their authority. They leave for flatter organizations.

What AI Changes

AI accelerates the case against layers:

- Information aggregation: AI can aggregate and synthesize information across the organization, replacing the information-rollup function of middle management - Decision support: AI can provide decision support at every level, reducing the need for layers to provide context and analysis - Quality assurance: AI can review outputs at scale, replacing the quality-control function of layer-by-layer review - Speed expectations: In an AI-enabled competitive environment, organizations that move slowly lose. Layers slow movement.

The organizations flattening are responding to these pressures. The layers that once enabled scale now prevent the speed that AI makes possible.

The Flattening Trap

There is a trap in flattening: removing layers without redesigning decision authority.

When you eliminate a layer of management, you eliminate the people who made certain decisions. If you do not explicitly reassign that decision authority, one of two things happens:

1. Decisions escalate: The decisions that the eliminated layer made now escalate to the layer above, overloading senior leaders 2. Decisions stall: No one knows who should make the decision, so it does not get made

The organizations successfully flattening are not just removing layers. They are redesigning decision authority at the same time — explicitly defining who now makes the decisions that the eliminated layer used to make.

What Successful Flattening Requires

1. Decision authority mapping: Before removing a layer, map every decision that layer made. After removing, explicitly assign each decision to someone else.

2. Span of control redesign: Managers who previously had 5 direct reports may now have 15. This requires different management approaches — more delegation, more trust, different oversight cadences.

3. Information system redesign: If layers provided information rollup, what replaces that function? AI-enabled dashboards, automated reporting, and real-time visibility tools.

4. Career path redesign: If layers provided career progression, what replaces that? New advancement paths that do not depend on management promotion.

5. Change management: Flattening creates anxiety. Workers who expected to become managers see that path disappear. The change requires deliberate communication and support.

The Organizations Leading

Citi, Telenor, Meta, and others are leading this shift. They are not waiting for the perfect plan. They are flattening now and learning as they go.

The organizations following will need to flatten too — or accept that their decision latency will be a competitive disadvantage against flatter, faster competitors.

The question is not whether to flatten. It is how to flatten without creating chaos.