Gallup's latest employee engagement report is out, and the headline number is the lowest in 15 years. In the third quarter of 2026, U.S. employee engagement fell to 29%, down two points from the previous quarter and its lowest level since 2011.
Each percentage point represents about 1.6 million workers, so that two-point drop amounts to some 3.2 million fewer engaged employees in a single quarter. Since engagement peaked in 2020, the total is closer to 11 million.
Engagement numbers drift for a lot of different and often competing reasons. This drop is unusual in how clearly it points to a culprit.
Where engagement is falling fastest
According to gallup, fully remote employees saw engagement fall five points in a single quarter; hybrid employees, two. On-site workers held roughly steady. But the sharpest declines came at organizations that have adopted AI, which is a clear reversal of what Gallup found only a few months ago. In July, AI looked like a tailwind: engagement ran higher at companies using it, and higher still where leaders explained their plan and managers helped people put it to work.
By Q3, for remote and hybrid employees, most of that advantage had vanished.
Responses to two of Gallup’s standard survey questions did much of the damage. The share of people who strongly agree that they know what's expected of them at work fell to 46%, down from 61% in 2015. And only 37% strongly agreed that a supervisor or someone at work cares about them as a person, ten points below where that figure stood in March 2020.
It would be easy to treat these as two separate problems, one about clarity and one about warmth.
To me, they read more like a single problem, just described from different angles.
What AI is doing to today’s workforce
AI changes work in ways that are easy to measure — and ways that aren't.
The measurable part is speed: a draft that took an afternoon now takes a few minutes. But what’s missing in that efficiency gain is everything under the hood: researching a problem, talking to experts or colleagues, building understanding of product or business context. Efficiency gains also mask what we once generally called "collaboration.” You know — things like asking for feedback from a peer, or handing something off to those with sharper perspective or skill.
With AI, we’re outsourcing those small, ordinary dependencies on other people that make a job feel like shared, well, work.
When those exchanges thin out, a few things go with them. People lose the informal cues that tell them whether they're doing the job well. And they lose the incidental contact that makes a workplace feel like a place with people in it.
Gallup puts it carefully, raising the possibility that physical distance becomes psychological distance. I disagree. I think it has nothing to do with the physical, but everything to do with cultural. For a remote employee whose most frequent collaborator is now about as close and available as you can get, the loneliness has never been more apparent.
None of this is an argument against AI. In Gallup's earlier analysis, organizations that paired frequent AI use with a clear plan and active manager support saw engagement levels reach 53%, a figure that would be remarkable in any year.
Technology isn't the problem. The problem is that AI removes the reasons people used to talk to each other, and very few companies have figured out how to deal with the cultural repercussions.
Building connection back in, deliberately
At Bonusly, we spend most of our time thinking about the moment one person tells another that their contributions mattered. We've come to see those moments less as a perk than as a kind of cultural foundation, because each time someone meaningfully shares recognition, it speaks to both of the questions Gallup's respondents are struggling with today.
Recognition tells someone, in specific terms, what good work looks like now, at a time when the definition keeps shifting. And it tells them that they are part of something with other people in it.
Gallup's own advice to leaders lands in a similar place: be clear about how AI applies to each role, help managers support people through the change (especially those leading remote and hybrid teams), and share real examples of the technology used well.
That last recommendation deserves underscoring. A public thank-you to the colleague who found a smart way to use a new tool is an example shared, an expectation made clear, and a small act of connection, all at once.
For leaders, none of this necessarily requires a net new way of doing things. It just requires remembering that building employee connections is part of the job, even as AI reorganizes everything around it. It’s on leaders to make time to talk about how the work is changing, to be specific about what good looks like, and to say so, out loud and in front of the team, when someone gets it right.
Gallup’s engagement numbers will move again. They always do. What's worth watching is whether organizations read this quarter as a blip, or as an early sign that work is becoming faster, more capable, and more solitary all at the same time.
The tools will keep getting better at the work. Making sure people don't do it alone is still very much up to us.
