AI Is Removing Management Layers. Who Does Their Work Now?

AI is making it easier to run organizations with fewer management layers. Reducing layers can improve speed, accountability, and cost-efficiency, but it can also expose gaps in decision-making and leadership support. How can leaders create direction without holding everything too tightly?

Leaner organizations are often designed to move faster. They promise less bureaucracy, shorter decision paths, closer connection to the work on the ground, and more direct accountability.

Yet fewer layers don’t automatically make an organization more effective. In some cases, they simply remove the buffers that once absorbed ambiguity. When organizations flatten and spans of control widen, leaders are often asked to carry more responsibility with fewer layers around them.

Moritz Drerup, Managing Partner at Signium in Hamburg, says,

“With less hierarchy to carry the load, leaner organizations need more deliberate leadership. Without clear priorities and distribution of accountability, leaders may face more direct reports, more information, and more decisions competing for their attention. Teams may be told they’re empowered, but still lack the decision rights or confidence to act without escalation. It quickly becomes overwhelming.”

AI is widening spans of control and changing how leadership works

For years, organizations have been working to simplify structures, reduce duplication, and bring decision-making closer to customers, markets, and operations. This shift is accelerating as digital tools and AI-enabled systems make it easier to track work and performance data and to connect leaders to activity across wider spans.

Gartner has predicted that, through 2026, 20% of organizations will use AI to flatten their organizational structure, eliminating more than half of current middle-management positions. This points to a wider pattern: organizations aren’t only reducing layers for cost reasons. They’re exploring how technology can support broader spans of control, faster information flow, and more efficient management models.

This means leaders have fewer intermediaries between themselves and the business’s day-to-day work. They often receive more direct signals from teams, customers, dashboards, and systems. They may be expected to respond faster, interpret more information, and hold more people accountable without the same level of management infrastructure beneath them.

When there are fewer layers, uncertainty can spread more quickly. When priorities are unclear, more people feel the impact. When decision rights are vague, more issues move upward. The organization may look leaner on paper, while leaders become more stretched in practice.

From an executive search perspective, this is also changing what companies ask for in senior leadership roles. Drerup sees this particularly in technology appointments:

“A few years ago, a CTO brief was often about depth in one domain. Today, it’s more about span: leading engineering, data, and AI adoption at once, often with fewer managers in between. The leaders who thrive aren’t simply managing more of the same. They’re able to stay close to the work while still setting direction.”

Where do future leaders learn to lead when AI removes the middle?

When companies reduce middle-management layers, they may also remove some of the roles where future leaders traditionally learned to lead.

Middle management has often been more than a coordination layer. It has been a training ground where people learn judgment, prioritization, and how to lead through others, but with stakes that still felt manageable. If that layer becomes much thinner, organizations need to think more intentionally about where the next generation of leaders will develop these capabilities.

AI adds a second pressure to the same pipeline. Much of the work that used to build judgment in early leadership roles, such as drafting the analysis, preparing the recommendation, and chasing down the missing information, is also the work that is easiest to automate. The layer where leaders were trained is thinning from above through delayering and from below through automation.

Drerup, who has spent more than ten years placing technology leaders, puts it plainly:

“If a first-line manager no longer builds the analysis themselves, they lose the practice that taught them what a good decision looks like. That is a slower problem than an open vacancy, and a harder one to correct.”

He sees the same pattern in the candidate market:

“As a search consultant, I often see this before it shows up formally in the business. The pool of leadership-ready candidates can start to thin when the training ground is rationalized away.”

This is already influencing how some companies think about talent. Drerup recently spoke with the CEO of an e-commerce company who was concerned about where the company’s future technology leaders would come from:

“Rather than wait for traditional career steps to produce them, the business is hiring strong junior, AI-fluent talent and building clearer roadmaps to help them grow into future leadership roles.”

That is what leadership development may need to look like in a flatter structure. When the old stepping stones are removed, organizations need to create new ones.

Intentional design: Who does the work when roles are removed?

One of the mistakes organizations make is assuming that fewer layers automatically mean more empowerment. In reality, empowerment depends on whether people have enough context, authority, capability, and trust to make good decisions.

In more traditional structures, layers often played a translating role. They filtered information, interpreted strategy, coordinated work, managed trade-offs, and escalated decisions gradually. Some of that may have been inefficient, but the work itself was still necessary.

When layers are removed, the coordination work doesn’t disappear. Drerup explains:

“If these roles aren’t assumed by someone, the leaner structure can create the opposite of what was intended. Instead of speed, there’s hesitation. Instead of accountability, there’s confusion. Instead of empowerment, there’s escalation and overwhelm. If a business is reducing layers, the leadership model needs to be part of the design from the start.”

AI changes which parts of that work a system can absorb. Reporting, status aggregation, and routine tracking can increasingly be automated. Judgment about trade-offs, the reading of context, and the work of aligning people across functions cannot. Organizations that flatten on the assumption that software will take over the whole of a layer’s contribution tend to notice the difference late.

Drerup adds:

“When companies remove a layer and hand the coordination to a system, they usually automate the visible half of the job. The half that made the layer valuable, knowing which trade-off to settle and which to escalate, stays behind, unassigned.”

Where should ownership sit? If the roles closer to operational reality don’t have sufficient authority to make appropriate decisions, the organization remains dependent on hierarchy even after it has been reduced. If they do have authority, leaders need to make sure they also have the context and boundaries to use it well.

AI shows leaders more. It does not decide what matters.

Leaders in leaner organizations often see more. They have access to more data, more direct feedback, more real-time reporting, and more frequent contact with teams across the business. This can be valuable, but it also creates a new test of judgment.

AI-enabled tools may help leaders spot patterns, automate reporting, or reduce the need for manual updates. Yet data can’t decide what matters most or fully explain context. It can’t always distinguish between a temporary variation and a deeper problem, nor can it determine whether a leader should intervene, coach, delegate, or step back.

Drerup elaborates:

“Leaders who can see more may feel tempted to interfere more. Dashboards can become a substitute for trust. Performance data can be used to micro-manage activity rather than focus attention. Over time, this weakens the ownership that leaner structures are meant to create. Leaders become overwhelmed while teams are unable to act. Ultimately, leaders become the bottleneck.”

Leaders in leaner organizations must use data to sharpen judgment, not replace it. They can use it to ask better questions, spot where teams may need support, and separate small fluctuations from issues that require action.

Traditional leadership habits become more exposed

Leaner organizations leave less room for vague leadership habits. Behaviors that may have been absorbed or corrected by layers in the past can become more damaging when fewer people sit between strategy and execution.

1. Leading through escalation

If too many decisions continue to move upward, a flatter structure becomes slower than the hierarchy it replaced. Leaders may become bottlenecks, while teams learn that accountability is encouraged in theory but limited in practice.

2. Communicating in broad themes

Senior leaders often talk about growth, transformation, innovation, efficiency, or customer focus. These themes may be directionally useful, but they are not always specific enough to guide action. In leaner structures, communication needs to reduce assumptions. People must understand what matters now, what can wait, where trade-offs should be made, and what outcomes they are expected to own.

3. Delegating tasks instead of assigning ownership

In a leaner organization, delegation can’t simply mean passing tasks down. It needs to mean transferring ownership within clear boundaries. People need to know not only what they are responsible for, but what they are allowed to decide.

4. Using visibility as a reason to micromanage

When leaders have access to more performance data, they may unintentionally pull decisions back toward themselves. This undermines the very purpose of a leaner model.

5. Assuming structure alone will change culture

Removing layers may signal a desire for speed and ownership, but people still need the confidence to act. That confidence comes from repeated clarity: clear priorities, clear decision rights, clear expectations, and clear consequences.

The balancing act: Leading without holding too tightly

Leaner organizations can be faster, more responsive, and often more cost-efficient. Realizing these benefits depends on how leadership evolves alongside the structure.

As Drerup puts it:

“In a leaner organization, effective leadership isn’t about holding everything more tightly, however tempting that may be. It’s about knowing where your involvement adds value and where it slows progress, and how to empower people with the confidence and resources they need to move forward. That’s the balancing act: being close enough to guide the work, while giving people enough room to own it.”

Although leaders must be clearer, more focused, and more confident when delegating, this is not only an individual challenge. Organizations must also make sure the structure supports the behavior they expect. If decision rights remain unclear, or if people are given more responsibility without enough context, leaner structures can create risk and friction rather than momentum.

Done well, fewer layers mean less distance: between a decision and the action that follows it, between leaders and the work itself, and between people and the room they need to grow. That is not something a structure delivers on its own. It is something leaders design, deliberately, and it is how the next generation of them is made.