If there’s one clear trend emerging in 2026, it is that long-feared AI-driven job cuts are beginning to land. That is bad news for employees generally, with companies such as Oracle carrying out multiple rounds of layoffs across offices and departments. But within that broader trend, middle managers appear to be bearing a disproportionate share of the impact.
Between May 2022 and May 2025, manager headcount at public companies fell 6.1%, compared with a 4.6% decline in executive roles, according to analysis reported by CNBC. Meta, Amazon, Google, Intel, and Estée Lauder are among the major employers flattening their hierarchies, presenting fewer layers as a route to faster decisions and lower operating costs. Companies are increasingly comfortable making these cuts because AI can now take on much of the coordination work that once justified large management structures, and Gartner predict that through 2026, one in five organizations will use AI to flatten their structures.
But reducing hierarchy is not necessarily the same as building a more effective organization. Stacy Parker, Co-Founder and Managing Director of Blu Ivy Group, told UC Today that it could have the opposite effect. Drawing on the company’s analysis of more than 60 client research reports on organizational performance and management, Parker explains why companies should be wary of treating middle management purely as an overhead.
AI Efficiencies or a Correction for Overhiring?
Amid all the talk of AI efficiencies driving layoffs, it is important to recognize the context that came before it. As Parker explains,
“Back in 2021, 2022, across the globe, companies were hiring at a pace that had never happened for organizations.”
As growth expectations tightened, those layers became an obvious target for finance leaders seeking a leaner cost base. With AI improving the ability to distribute information, monitor work, and generate management reporting at speed, all roles of management fulfill, they became the focus of those cuts.
Parker identifies reporting, meeting preparation, project updates, and real-time alerts as areas where AI can already make a material difference. “They’re doing that extremely well,” she says of the technology’s ability to handle these routine tasks. But those activities are not the full value of a manager. The danger begins when an organization treats the ability to automate them as evidence that it can remove the people responsible for context, escalation, and judgment.
That matters because the wider organizational picture is already fragile. In Blu Ivy Group’s review of more than 67 companies, more than 87% identified leadership clarity and communication among their main challenges. The research does not attribute those problems to flatter structures. However, it suggests businesses considering deeper management cuts may be removing a layer of support when employees and customers already need clearer direction, accessible leaders, and effective channels for feedback.
What Can Businesses Do About It?
With that in mind, Parker argues that companies need to rethink how they translate AI efficiencies into organizational change. “Let’s not start with how much headcount can we reduce,” Parker says.
“Let’s start with how can we unlock greater productivity and performance and connection to our markets that we serve.”
That reframes AI transformation as work redesign rather than a cost-cutting exercise, forcing leaders to identify which tasks can safely be automated and which decisions need human ownership.
Parker’s anecdotal evidence from organizations taking that approach was encouraging. Companies that retained management while shifting some responsibilities toward coaching, reskilling, and AI-enhanced projects with employees showed greater performance, stronger organizational coherence, and faster scaling. The point is not that every management role should be retained unchanged, but that the layer’s value may increasingly lie in how it develops people and supports new ways of working.
For the managers who remain, the role should move away from administrative control and toward coaching, capability building, and judgment. Parker argues that the human contribution will increasingly involve “interpreting what AI output is and what needs to be questioned and dug into.” Rather than spending their time producing updates or routing information, managers should be expected to coach employees, develop skills, and intervene when AI-generated recommendations require challenge or context. That makes the role less about administrative oversight and more about helping teams use AI productively without outsourcing judgment to it.
That is also an operational requirement for agentic AI, not simply a people initiative. Microsoft’s research calls this emerging role the “agent boss”: a person who sets direction for digital workers, evaluates their outputs, and resolves exceptions. Businesses need governance that assigns accountability for agent decisions, data quality, and customer impact. Without it, a flatter model can merely transfer coordination work from an eliminated layer to already stretched specialists and executives.
Why Managers Still Matter
The question is not whether the traditional middle-management job will survive unchanged. It will not. Routine coordination and information routing are increasingly automatable, and organizations that retain those tasks as the center of management work will carry unnecessary cost and friction. The real question is whether companies can eliminate middle management without creating new problems, or whether they must redesign the role for a workplace where managers are less responsible for administration and more responsible for judgment, coaching, and directing human-AI teams.
As AI agents move from pilots into everyday workflows, the immediate challenge is to distinguish between work that can be automated and accountability that cannot. A dashboard may show a performance problem, and an AI agent may summarize the likely causes, but someone still has to decide what matters, test the assumptions behind the output, and take responsibility for the response. In a flatter structure, that responsibility will sit with fewer people, making the quality of those remaining managers more consequential, not less.
The companies that get this right will not be those that preserve an outdated hierarchy, but those that use AI to remove low-value management work while retaining human accountability where it matters.
