The Flattening Middle: Protecting Your Leadership Pipeline as Entry-Level Roles Shrink

Ashley Hunt Founder & Director, Veraison

For most of the last century, the path to senior leadership looked reassuringly predictable. The organisational pyramid wasn't just a structure for getting work done — it was a development engine.

People joined in an entry-level role, learned the business from the ground up, took on a small team, then a larger one, and somewhere along the way absorbed the judgement, relationships and instincts that senior leadership demands. Every rung was a place to learn before the stakes got higher.

That engine is now being quietly dismantled, and most organisations haven't noticed.

What's actually flattening

As AI moves from experiment to everyday infrastructure, the work that once defined junior and middle roles is the work being automated first. Routine analysis, first drafts, scheduling, data entry, basic coordination — the very tasks that gave early-career employees something to cut their teeth on — are increasingly handled by software. Industry analysts describe organisations flattening their structures, automating tasks, and reducing the number of entry-level roles, all at the same time.

The shift isn't hypothetical. According to Korn Ferry, more than half of talent leaders plan to add autonomous AI agents to their teams in 2026. Workforce researchers warn that as entry-level and middle-management roles shrink, remaining managers are expected to move toward more strategic, higher-value work — while organisations face a quieter, slower-burning problem: how do you maintain a leadership pipeline when you've removed the rungs people used to climb?

Part of what's driving the cuts is a shift in what managers believe is possible. Year on year, more of them now see roles on their own teams as replaceable.

The World Economic Forum estimates that 39% of workers' core skills will change by 2030. We are not simply doing the same jobs with fewer people. We are changing what the jobs are — and in doing so, changing how, and whether, people learn to lead.

Why this breaks the pipeline

A leadership pipeline is really a sequence of low-stakes opportunities to practise high-stakes skills. You learn to give difficult feedback on a team of three before you're responsible for three hundred. You learn to read a room, sit with ambiguity, and recover from a poor decision while the consequences are still survivable.

The middle — where leaders are made — gets hollowed out.

When organisations automate away the early roles, they don't just lose headcount. They lose the apprenticeship. The danger is a workforce that is suddenly bimodal: a layer of senior leaders who came up through the old system, and a layer of capable individual contributors who have never been given the small, formative leadership experiences that used to happen naturally.

This is why leadership development has become, and remains, a top priority for people leaders heading into 2026. It is also why the organisations that take it seriously tend to pull ahead. Research from McLean & Company found that organisations with strong leadership are more than twice as likely to excel at innovation. Capability at the top is not a soft nicety; it is increasingly the thing that separates organisations that adapt from those that stall.

What leaders can do now

The instinct to let the pyramid flatten and "see what happens" is understandable, especially when AI is delivering real efficiency gains. But the cost of inaction shows up years later, as a sudden, expensive scramble for leaders who were never developed. A more deliberate approach involves a few shifts.

Make development intentional, not incidental. When growth no longer happens automatically as a by-product of climbing the ladder, it has to be designed in. That means building explicit opportunities for early-career people to lead a project, chair a meeting, own a relationship, or carry a decision — even when AI could technically do the underlying task faster. Some inefficiency is an investment in capability.

Redesign roles around uniquely human work. As AI absorbs routine tasks, the value of human judgement, ethical reasoning, conflict resolution and change leadership rises sharply. These are precisely the capabilities that don't automate — and precisely the ones future leaders need most. Reshape early roles so that people spend their time on the judgement-heavy work, not in competition with the software for the routine work.

Treat managers as developers of people, not just managers of output. With fewer layers, the managers who remain carry more responsibility for growing those beneath them. That only works if they are equipped and rewarded for it. Many managers were never trained to recognise potential or coach it forward; in a flatter organisation, that capability gap becomes a pipeline risk.

Widen the definition of a leadership pathway. The linear ladder is fading, but leadership can grow along other lines — through cross-functional projects, rotations, mentoring, and stretch assignments that build breadth rather than just seniority. The organisations that thrive will be those that create many small doorways into leadership rather than one narrowing staircase.

Taken together, these are four deliberate moves — a way to keep developing leaders once the structure stops doing it for you.

The window is open now

The flattening of the middle is not a distant trend. It is happening in the structure of teams being designed this year. The organisations that will have strong leaders in five years are the ones making deliberate choices today — choosing to develop people on purpose, because the old system that did it for them is going away.

AI can take a great deal of work off our plates. What it cannot do is grow the next generation of leaders. That remains, stubbornly and reassuringly, human work — and it is work worth protecting.