AI Impact on Middle Managers: What the Data Shows

The AI impact on middle managers is no longer a forecast. Over the past year, big employers have cut management layers and named AI as a reason. Meanwhile, fresh surveys show the managers who remain are using AI themselves, and they trust it less than their bosses do. This explainer pulls the data together.

For years, the fear was that robots would replace factory workers. However, the first big cuts landed on desks. In October 2025, Amazon said it would remove about 14,000 corporate jobs, according to Fortune. That was roughly 4% of its white-collar staff.

The memo came from Beth Galetti, Amazon’s senior vice president of people experience. It said the company wanted to reduce bureaucracy and remove organizational layers. In addition, CEO Andy Jassy has said generative AI will take on planning, analytics and forecasting work.

Those tasks sit at the heart of a middle manager’s week. So when software handles them, the layer above the team starts to look thin.

Amazon was not alone. Ramp tracked the pattern in a May 2026 analysis, and Fortune covered the earlier wave. The table shows who did what.

The language repeats across these firms. Coinbase CEO Brian Armstrong said the company would no longer have pure managers. Similarly, Block’s Jack Dorsey endorsed a player-coach model, where leaders also do hands-on work.

We saw the same logic in our report on the Oracle workforce reduction. Money moves to AI, and headcount pays for it.

The squeeze shows up in the numbers too. According to Gallup figures cited by Ramp, the average manager had 10.9 direct reports in 2024. A year later, that number was 12.1.

Furthermore, a Korn Ferry survey from 2025 found that 41% of professionals saw manager roles cut at their company. The side effects were clear. In the same survey, 37% said they felt directionless, and 43% said their leaders lacked alignment.

The trend also predates the chatbot boom. A Bloomberg and Live Data Technologies analysis found that middle managers made up about a third of layoffs in 2023. Therefore, AI is speeding up a shift that had already begun.

Gartner put a number on where it leads. It estimated that by 2026, one in five organizations would use AI to eliminate at least half of their management layers.

Not entirely, but it takes the routine half of the job. Fortune describes the bet plainly. Algorithms can handle coordination, reporting and some decisions that human managers once owned.

Managers already face that question about their own teams. HR Dive reported a ResumeTemplates.com survey of 1,000 U.S. managers who use AI at work. One in five had been asked to decide whether AI could do a job.

Their answers were blunt. Among managers asked to judge a specific role, 75% said AI could replace it. Still, the sample only covers managers who use AI, so it does not speak for everyone.

The same survey shows managers leaning on AI for hard calls. In fact, 59% said they use AI when deciding who to lay off, and 58% use it for firing decisions.

Most keep a hand on the wheel. For example, 57% said they would never let AI decide without supervision, and 91% would override a recommendation they disagreed with. However, 17% said they go hands-off often or all the time.

The guard rails look weak. According to the survey, 38% of managers have never had training on the ethical use of AI in HR decisions. In addition, 58% could not confirm that their company tested the tool for bias.

New research published on October 7, 2026 adds another angle. TeamViewer found that middle managers want more human control over AI than executives do.

The gap is wide. When AI resolves IT issues that touch security, data or operations, 46% of middle managers want a human involved. Only 24% of executive leaders said the same. Likewise, 43% of middle managers said stronger security and privacy protections would raise their comfort, against 34% of executives.

Sapio Research ran the study in April and May 2026. It surveyed 4,200 people across nine markets, including the U.S., the U.K., Germany and Japan. Overall, 73% of workers said they trust AI more when it shows what it is doing and why.

Accountability remains murky, though. Only 49% said it is clear who answers when an autonomous system makes a mistake. TeamViewer CISO Jan Bee argued that security has to come before autonomy, not after it.

First, keep the scale in view. Challenger, Gray & Christmas data cited by Ramp shows total corporate layoffs down 10% this year. Tech layoffs, however, are up 33%. So the pain is real but concentrated.

Second, the role is changing more than it is vanishing. Companies now want managers who coach, build and ship alongside their teams. As a result, pure status reporting is the riskiest work to own.

Finally, the people in the middle carry the risk when AI goes wrong. The TeamViewer numbers suggest they know it. Because of this, managers who learn to supervise AI well may become harder to cut, not easier.

To see which coordination tasks software already handles, browse our guide to the best AI workflow automation tools. If meetings eat your week, compare the best AI note taker apps we tested.

Sources

Fortune, Ramp, HR Dive, TeamViewer (PR Newswire). *Photos: Working in open office space by Crew, CC0, and Amazon Spheres 2018 by Andy Li, CC0, both cropped and resized.*