A surge in workplace use of generative AI is failing to deliver clear returns, and a growing share of employees say the tools are slowing them down instead of speeding them up. New research from BetterUp Labs and Stanford reports that many companies are stuck with “workslop,” the polished but shallow output that pushes extra effort onto coworkers. The findings land as leaders search for a path to real productivity gains without eroding trust or quality.
The study points to a common pattern. Organizations roll out AI broadly, but offer little guidance on when and how to use it. Workers then spend time fixing or rebuilding AI-generated content. The result is a hit to productivity and team morale at the very moment leaders expect efficiency.
Background: A Fast Uptake, Thin Payoff
Generative AI tools have spread quickly across teams that write, code, and analyze data. Many firms hoped for measurable gains in speed and accuracy. Yet the research suggests that these benefits are uneven. Employees report frequent encounters with surface-level content that looks refined but lacks depth.
“Workslop—content that appears polished but lacks real substance, offloading cognitive labor onto coworkers.”
According to the study, 41% of workers have dealt with this kind of output. Each instance costs nearly two hours of rework. Those delays compound across projects and departments, creating tension and skepticism about AI.
Inside the ‘Workslop’ Problem
The core issue is not only accuracy. It is fit for purpose. AI can draft emails or summaries in seconds, but teams still need judgement, context, and domain knowledge. When those are missing, the output forces others to think through what the tool should have done in the first place.
Researchers warn that poor norms drive poor outcomes. Vague mandates to “use AI” push people to apply it everywhere, even when a task needs original analysis or direct expertise.
“Leaders need to consider how they may be encouraging indiscriminate organizational mandates and offering too little guidance on quality standards.”
The fallout shows up in three areas: productivity, trust, and collaboration. Time lost to fixes hurts delivery. Peers grow wary of AI-generated drafts. Teams spend more energy checking each other’s work rather than building on it.
How Leaders Can Course-Correct
The report outlines a practical playbook. It starts with modeling good use. When leaders show where AI adds value—and where it does not—teams follow suit.
- Set clear quality standards and review steps for AI-assisted work.
- Define use cases where AI is encouraged, optional, or out of bounds.
- Share examples of strong prompts and effective edits.
Another theme is culture. The researchers urge leaders to create a “pilot mindset.” That means structured trials, small bets, and quick feedback. Employees should feel high agency and optimism, but also be accountable for results.
“To counteract workslop, leaders should model purposeful AI use, establish clear norms, and encourage a ‘pilot mindset’—promoting AI as a collaborative tool, not a shortcut.”
Balancing Speed With Substance
Teams that win with AI pair speed with rigor. AI drafts can jump-start a task, but users still need to set the goal, supply context, and verify claims. Review checklists and peer walkthroughs help. So does tracking rework time, which reveals where tools help and where they hinder.
Some functions may see faster gains than others. Routine formatting, code scaffolding, and summarizing long documents are strong candidates. Original analysis, sensitive communications, and final decisions should remain human-led, with AI as support.
What to Watch Next
As executives refine their playbooks, the measure that matters is net time saved without loss of quality. Companies can pilot AI on narrow workflows, define success metrics, and expand only when rework declines.
The study’s message is clear. Mandates without guardrails invite workslop. Purposeful use, shared norms, and a test-and-learn approach can turn AI from a shortcut into a partner. Leaders who set that tone may finally see the returns they expected.
