The Era of Zombie Employees

Output is up. Attachment is gone. Both are true, and in the same workforce at the same time they are not supposed to be.

Gallup put employee engagement at 20 percent in 2025, its lowest reading since the pandemic and a second straight year of decline. Flip that figure and roughly four in five workers call themselves not engaged, a drift Gallup values at around $10 trillion in lost output worldwide, close to nine percent of global GDP. Over the same period, US labor productivity, output per hour, rose 2.2 percent in the year to the second quarter of 2026, after a 2025 strong enough that economists at the Federal Reserve Bank of San Francisco put the odds at 57 percent that the economy had entered a faster growth gear. On paper the American worker has rarely looked more efficient. Off paper, rarely more absent.

The gap between those two numbers has been named again and again over the past two years, each label catching a little more of the same retreat: quiet quitting, then quiet cracking, then job hugging, then Gallup's own Great Detachment. The bluntest name is the zombie employee. These are not slackers, and they are past the performative minimum of quiet quitting, where the point was to be seen doing less. They take the assignment, agree to the deadline, and let it die quietly in a meeting three weeks out. The work still gets done. The body still clocks in. Whatever part of them used to care about the job is somewhere else.

Detachment on that scale should be dragging output down. It is not.

The work stopped running through the worker

Output keeps holding because it no longer runs through the worker the way it used to. Teams got cut hard through late 2023 and 2024 and were then handed the same targets, and the gap a checked-out person once left behind now gets filled by software. The report still gets assembled. The support ticket still gets closed. None of it depends anymore on somebody wanting to do it.

Picture the person who used to write that report. Two years ago a bad week from them showed up downstream, in a number nobody chased or a client who noticed the difference. Today that person opens a model, pastes in the inputs, edits what comes back, and sends it. The report is fine. It is always fine now. Whether they cared this week leaves no mark on the output. Multiply that across a department and you have a productivity line that has quietly stopped reporting on the state of the people beneath it.

The obvious objection is that none of this has anything to do with engagement. Cut a team and hold output flat, and output per hour climbs on its own, because you have dropped the least productive hours and left the rest to the survivors. But that only buys a single step up. It explains a jump in the quarter after the layoffs, not output holding steady while the workforce keeps detaching year after year. The gain from a smaller headcount fades once that headcount becomes the baseline. What we are watching is a level that stays put while the thing that used to sustain it, people caring whether the work was good, keeps draining away. The tools are absorbing the difference.

That rewires what a job even is. For most of working history, the extra effort a motivated person brought was the difference between a company that functioned and one that stalled. Engagement counted because the work depended on it. Once that dependency breaks, the chart stops tracking morale, and a productivity line can drift upward over a workforce that has quietly checked out.

The workers did the math

The workers have read the situation correctly. What looks like an attitude problem is closer to a calculation.

Start with what they say out loud. In a June 2026 survey by Founder Reports, 36 percent of US employees said going above and beyond would not protect their job, so they saw little reason to do it. A 2025 Harris Poll for Express Employment found that 73 percent of job seekers thought no job was safe regardless of performance, and two-thirds of people who once felt secure said they no longer did. Call it cynicism if you like, but it tracks the incentive they have been handed.

They can see why, because they can see the layoffs. Challenger, Gray & Christmas counted 101,743 job cuts in the first half of 2026 that companies blamed on AI, close to double the 54,836 the firm logged across all of 2025. Employees are watching money leave payroll and move into the systems now covering their absence. Under those conditions, holding something back is simply the reasonable read.

It gets harder to argue with once they check their share of what they produce. As productivity rose, labor's share of business output fell to 52.8 percent, the lowest in a record that begins in 1947. Output per hour is up while the slice of it returning to the people making it sits at a three-quarter-century low. Line the facts up. Effort will not save the job. The job may go to software regardless. Whatever value does get made is increasingly kept by someone else. A system built to teach workers that effort no longer buys safety would look a lot like this one, and they have learned the lesson.

Gallup already has a name for this

They stay because the exits closed. Gallup calls the mood the Great Detachment: a workforce that feels stuck rather than restless, unable to leave because the job market cooled and unwilling to bet on a future it does not trust. Late last year only 28 percent of US workers thought it was a good time to find a quality job, down from 70 percent in the middle of 2022. One force drains the will to push; the other locks the door on the way out. You will not find many zombie employees in a booming labor market. They surface when the exits narrow and the point of the work blurs at the same time.

What is left is the quiet opposite of the Great Resignation. Same discontent, none of the churn. People with every reason to walk and no safe way to do it, so they stay and go silent.

You can run on this for a while

The zombie employee, competent and entirely absent, is the visible end of all this. The frightening part of the AI transition was never only the people who lost their jobs. It was the message the layoffs sent to everyone still holding one. Your effort has been separated from your security. Your output has been separated from your enthusiasm. The operation will run about the same whether you are present or not. People are answering the way most of us would. They are keeping something back.

None of this reads as a crisis on a productivity chart, which is exactly why it is easy to miss. The output looks fine. The cost lands in the places the charts do not measure, like the judgment nobody bothers to sharpen anymore and the loyalty that used to catch the mistakes automation still misses. A company can run on checked-out people and capable tools for a good while. The trouble is that from inside that arrangement, nobody is especially motivated to notice when it stops working.