Beyond workslop: How AI is shaping workplace communication

Emails chock full of bullets and emojis, “delving” and “pivotal” galore—you’re not imagining it—your coworker has hit you with “workslop.”

Stanford professor Jeff Hancock and BetterUp Chief Scientist Kate Niederhoffer were on a team that coined the evocative term in fall 2025, just before Merriam-Webster chose “slop” as the dictionary’s word of the year.

And don’t expect orderly inboxes or chat windows any time soon, they said. “We’re going to see, potentially, a worsening, because it takes a long time to change these organizations,” Niederhoffer told Morning Brew.

AI is everywhere in the office now, and it’s brought with it a rambling lexical flatness that can vex your coworkers when inserted into emails, memos, and reports. The technology has already begun to reshape in-office communication and information flows, experts said.

And with enterprise software companies pitching AI agents not just as workplace tools but as fellow nodes on the org chart, HR leaders are going to have to rethink the norms of how workers communicate with each other—and with bot intermediaries—to avoid a workslop overload.

“It’s a huge, huge question right now, across many dimensions,” Gabriella Rosen Kellerman, who worked on the workslop research in her previous C-suite role at BetterUp and is now a BCG director, told us.

Mistrust missed

After they published their original research on the topic, Niederhoffer and Hancock felt that executives responding to it tended to overly fixate on the productivity angle. Yes, workslop can waste everybody’s time and add more busywork. But a bigger danger is that thoughtless AI communications can cause teams to look down on and mistrust one another, Niederhoffer said.

“We felt like it was just misplaced attention there, and only there, when this other stuff, the negative cascade that happens throughout the fabric of the organization, is much more insidious and will have much more long-term damage,” she said.

At a time when surveys like the annual Edelman Trust Barometer show declining trust in institutions, Hancock told us that the workplace has remained one venue where “trust seems to be bridgeable.” Sloppy AI communication threatens that.

“It’s undermining the way teams work,” he said. “It’s undermining how we think about one another, and so there’s probably an even bigger hidden cost to the trust that is really crucial in organizations.”

Necessary friction

AI’s well-documented tendency toward sycophancy can also reinforce, rather than constructively challenge, problematic ideas or habits, according to Matt Poepsel, VP at hiring platform Predictive Index. As workers turn to AI for feedback on tasks like sending emails, AI is more likely to try to “sidestep the tension,” whereas a human giving feedback “could lean on the relationship, the body language, the interactive style” to tactfully deliver a criticism.

Skillsoft Chief People Officer Ciara Harrington said she hasn’t seen workslop become a problem in her company’s office yet, but she worries about it as workers become more AI-native. Skillsoft has been training its employees to think of AI more for feedback than as an email ghostwriter.

“As we’ve moved into this next phase, it’s really treating AI more like a coworker,” Harrington said. “You might ask [AI], ‘What would my CEO criticize about this? What would my team members criticize about this? What would they think is good about it?’”

Given the choice

Brian Jabarian is an economist at the University of Chicago’s Booth School of Business who studies how AI impacts information flows at work. He said one problem with companies imposing AI on workers for certain tasks is that people don’t work in tidy sequential processes. Brainstorming larger ideas might happen in the course of typing out emails, for instance.

“There is this parallel stacking of tasks…that we humans do that an AI doesn’t do. That can actually create a lot of mess when you deploy AI for one task because you didn’t realize that it was a bundle,” Jabarian said.

One key is to allow people to make a choice about when and when not to use AI, he said. “I use a lot of my emails not to communicate to friends, but to really think with them. So if you want me to use AI for writing my emails, you’re actually not helping me write my email—you’re taking away my thinking process,” Jabarian said.

Agentic future

Tech companies imagine a workplace of the future populated by a lot more AI peers—or at least direct reports. Microsoft has laid out a vision in which every worker is an “agent boss.” “People will manage teams of agents to do very complex things,” OpenAI CEO Sam Altman posted on X last week in introducing OpenAI’s new agent management platform, Frontier.

As more of these agents are introduced, all kinds of new questions about workplace communication could crop up, experts said. How do you design systems where people are accountable for their agents’ communications? What kind of psychological toll might interacting with AI agents all day take on a human worker?

Kellerman said BCG is currently studying many of these questions for a forthcoming report.

“[We’re] looking at how…managing lots and lots of agents affects lots of aspects of how we’re able to do our work, including social interactions, mental fatigue, and cognitive depletion. And it’s very significant,” Kellerman said. “These tools require a good deal of oversight…How do you not only oversee what you yourself are passing on to colleagues, but what the agents that you oversee are passing on to colleagues, and what’s your accountability for your agents’ work products?”

“It’s a very important step for us to start to understand that they are extensions of us in a way, and to take ownership of that. There’s also a question of, how much can one person’s brain hold? How much responsibility, accountability for how many of these tools at one time?”