Workslop: Why AI performance depends on how we think, talk and lead

Artificial intelligence (AI) was supposed to make us faster. In many ways, it has. A new report from The Wharton School of the University of Pennsylvania finds that many companies now see gains in productivity and performance. Yet those gains remain uneven across teams and industries, reflecting not what AI can do but how people use it differently.

Many leaders question the return. The same studies that celebrate AI’s promise also reveal that measurable impact is inconsistent. Our research points to a growing culprit: “workslop” – polished but hollow AI output that looks finished yet lacks substance, context or accuracy.

In a 10,000-person organization, that rework adds up: roughly two hours a day spent fixing unhelpful AI output equates to about $9 million a year in lost productivity.

The issue isn’t the technology; it’s the habits and behaviours we bring to it. According to BetterUp's research, employees trained in relational skills – listening, asking questions, providing context – interact 30% more with AI tools and produce higher-quality work.

Productivity now hinges on the quality of dialogue between people, and between people and machines. Three forces shape that dialogue: mindset, skills and leadership communication. Together, they determine whether AI is an advantage or a distraction.

Across our research, two clear mindsets emerge.

"Pilots" score high in agency and optimism, using AI deliberately to extend creativity and insight. They approach every interaction with colleagues and machines as part of the work itself, not as a shortcut around it. "Passengers" score lower in both. They are more pessimistic and fearful, and they rely on AI as a shortcut or avoid it altogether.

Agency and optimism give people confidence to navigate change. When these fade, dialogue falters. People stop asking questions, stop pushing for clarity and start letting automation do their thinking. The critical exchange that makes AI useful disappears.

Pilots approach conversation with both colleagues and AI as part of the job itself. That small distinction explains much of the performance gap across teams and leads to the next driver that shapes dialogue: skills.

In one BetterUp Labs study on generative AI (GenAI), people trained in relational skills – asking better questions, giving context and critiquing responses – engaged more effectively with AI and produced higher-quality work.

When people treat AI as a collaborator, the quality shift is immediate. Picture a marketing executive using AI to analyze customer feedback for a global campaign. The summary looks clean but it misses a key insight: customers trust the brand but doubt its follow-through.

She asks the AI to show the underlying comments, pinpoint scepticism and reframe the story from the customer’s view. The next draft captures the tension and the truth.

That courageous communication, surfacing assumptions and naming uncertainty, builds quality and trust. The most productive users don’t accept the first idea that reads well. They push the tool to think with them.

These aren’t technical skills. They’re deeply human. The same behaviours that strengthen human collaboration make AI infinitely more useful.

Leaders set the tone for how teams use AI and how readily it’s adopted.

Our research shows that requiring AI use increases the odds of developing a pilot mindset sixfold. However, when employees are satisfied with leadership communication that connects AI use to purpose and confidence, that likelihood rises to 21 times as likely.

But purpose alone isn’t enough. Participative communication matters as much. When leaders invite questions, surface uncertainty and act on employee feedback, they create the clarity and trust that reduces workslop.

When people understand how AI fits their role, when to rely on it and when to apply judgment, their work becomes clearer, faster and more relevant.

Intentional leadership communication turns compliance into collaboration. It turns anxiety into agency. And it directly improves both optimism and AI outcomes across the organization.

Without the human elements of mindset, skills and communication, AI multiplies mediocre work. The problem starts with technology but spreads through culture: overloaded teams, rigid mandates and unclear communication about AI create a productivity drag that no amount of automation can fix.

Workslop is a symptom. The underlying issue? Teams lack the human capabilities required to work effectively in an AI-augmented world. The future belongs to organizations that strengthen what technology cannot replace.

Our research shows that sustained development of these human capabilities can raise team performance by 50%, increase innovation by 90%, improve productivity by 15% and cut burnout in half. These gains don’t come from better algorithms. They come from investing in human transformation in three ways:

To refocus the workforce, top leaders will:

Effective reskilling includes the following initiatives:

To mainstream continuous development, business leaders can:

Leaders who act now will turn potential into performance. The future belongs to teams that learn to think together – in a world where their partners are inevitably both human and machine.