Glean: 6 Hours Lost Each Week Fixing AI

Workers Spend Over 6 Hours a Week 'Supervising' AI

A recent report from Glean highlights a concerning trend: workers are spending nearly an entire day each week overseeing artificial intelligence. According to the document, employees spend an average of 6.4 hours per week on what is referred to as "botsitting." This activity involves providing context to the AI, correcting its errors, and cleaning up the results produced by these systems.

Researchers emphasize that this situation is prompting many employees to consider changing jobs. While AI is supposed to lighten their workload, some find themselves spending long hours rectifying mistakes made by these technologies.

A report from the Work AI Institute at Glean, developed in collaboration with researchers from prestigious institutions such as Notre Dame, Stanford, and UC Berkeley, reveals that office workers spend an average of 6.4 hours per week "botsitting" AI. This term, coined by the authors of the report, describes the often-underestimated work necessary to make AI truly useful. The study surveyed 6,000 full-time workers in the United States, the United Kingdom, and Australia, primarily using digital tools between December 2025 and January 2026.

A "Often Boring" and "Exhausting" Job

The report's findings highlight a growing disconnect between individual productivity gains and overall company performance. This productivity paradox is a challenge many organizations face. While 87% of surveyed workers claim to use AI in their work and 75% believe it enhances their productivity, only 13% think their organization is experiencing a significant improvement because of it.

According to Rebecca Hinds, head of the Work AI Institute at Glean and co-author of the report, a significant portion of lost productivity is absorbed by unexpected tasks for employees. In the "Cognitive Revolution" podcast, Hinds described botsitting as a job that is "often boring" and "exhausting," which is "neither recognized, nor rewarded, nor tracked, nor measured, and certainly not incentivized within the organization."

The Risk of Departure

This additional workload appears to have a significant impact on employee morale. The report indicates that workers spending an abnormally high proportion of their time supervising AI are 73% more likely to actively seek another job.

"Workers who absorb this without recognition or reward become exhausted. Then they become bitter. After that, they start polishing their resumes," the report notes.

Employee frustration is not limited to just extra work. Many are now spending their time transferring information between disconnected AI systems, correcting errors, and providing context that the tools should already possess—thus becoming intermediaries for technologies that do not work well together.

In some cases, workers are also being asked to automate the parts of their jobs they enjoy the most. Hinds mentioned in the podcast that this particularly affects customer service employees, who enjoy building relationships but are increasingly expected to supervise AI agents instead. "This is what brings you joy and meaning at work," she said. "It's very dangerous."

Breaking the Botsitting Cycle

Researchers assert that the solution does not lie simply in deploying more AI. Organizations that achieve the greatest gains are often those that invest more in the work surrounding AI—helping employees access the right context, teaching them how to use the technology effectively, and establishing clearer standards for what good AI-assisted work looks like.

"The companies that are getting ahead are doing something different," the report states. "They are not spending a greater share of their AI time using AI. They are spending a greater share on the work that surrounds it: defining context, defining what 'good' looks like, developing judgment, and deciding what should never have been entrusted to a model in the first place."

The authors warn that the alternative is to continue paying the price of botsitting and "in the constant departure of people who are fed up with cleaning up after the bots."