AI saves workers 11 hours, but botsitting eats gains

Glean's Work AI Institute has published a survey of 6,000 digital workers on the effect of artificial intelligence in the workplace. It found that most respondents use AI at work.

Based on workers in the US, UK and Australia, the research found that employees believe AI saves them about 11 hours a week. But much of that gain is eroded by time spent checking outputs, fixing errors and moving between different tools.

According to the findings, 87% of digital workers use AI at work and 75% say it makes them more productive. Yet only 13% say AI has significantly improved their organisation's performance, pointing to a gap between personal efficiency gains and broader business results.

The report identified what it called "botsitting" as a major source of lost time. Workers said they spend 6.4 hours a week, or 37% of their AI-related time, feeding systems with context, supervising results, debugging mistakes, cleaning up generated work and switching between tools.

That was slightly higher than the 36% of AI-related time respondents said they spend using AI to complete work. The findings suggest that oversight and correction now take up at least as much time as direct use.

Hidden labour

The survey also pointed to risks around the quality of AI-assisted work. It found that 69% of AI users admitted to what the report termed "botshitting", including 41% who said they sometimes deliver work they could not explain if asked and 28% who said they had blamed AI for mistakes they caused.

Researchers said this raises questions for employers introducing AI tools into everyday processes without clear controls or training. The report argued that headline usage figures can hide how much human effort is still needed to make the systems usable.

Workers also appeared increasingly willing to treat AI as a colleague rather than just a software tool. The study found that 61% said AI helps them more with day-to-day work than their manager does, while 52% said it is easier to collaborate with AI than with human co-workers.

More than half of respondents, 55%, said they had sent a digital twin to a meeting on their behalf. Another 29% said they were comfortable with AI being involved in firing decisions.

Tool sprawl

The report found that many employees use multiple AI products to complete the same task. Among AI users, 77% said they juggle multiple AI tools each week, 33% said they use four or more, and 60% said they rerun the same prompt across several tools because the first answer was not good enough.

That pattern reflects a broader problem with access to information. The survey found that 53% of workers said important information they need is not accessible from their AI tools, making it harder to generate reliable outputs from a single system.

Employees using what the report described as context-poor AI were markedly more likely to report negative effects. Half said they felt worn out by AI, compared with 18% among those with better access to context. And 35% said they clean up after AI at least weekly, compared with 24%.

The same group was also more likely to say they deliver AI work they cannot explain, at 54% versus 26%, and to use unapproved AI tools, at 53% versus 21%. The findings suggest weak integration between AI systems and company information may be driving both fatigue and compliance risks.

Management challenge

The report found that workers who gain the most from AI are often the most likely to bend rules around its use. Among respondents described as high AI achievers, 54% said they use unapproved tools or use approved tools in non-compliant ways, 38% said they downplay AI's help to their manager, and 36% said they hide how much AI helps them.

It also suggested that some companies are adapting more effectively than others. In organisations described as transformative, 90% of workers said their employer provides enough AI training and support, compared with 52% elsewhere. And 84% said their employer formally rewards AI skills, compared with 48% at other organisations.

Those organisations were also more likely to treat AI as a chance to redesign work rather than simply add another tool. The survey found that 90% of workers in those businesses saw AI as part of a broader rethink of how work gets done, compared with 54% in other organisations.

Dr Rebecca Hinds, Head of the Work AI Institute, said the findings show many companies are focusing too heavily on AI adoption metrics. "Too many companies are treating AI adoption like a vanity metric - more seats, more prompts, more usage," said Dr Rebecca Hinds, Head of the Work AI Institute at Glean. "But adoption alone doesn't equal transformation. If employees are spending the productivity dividend on botsitting and botshitting, companies haven't eliminated work - they've created a new layer of overhead. The organizations that win will be the ones that ground AI in real enterprise context, apply the right guardrails, and measure success by business outcomes, not activity."

Hinds said the next stage of workplace AI will depend less on buying more tools than on changing how organisations manage them. "The next phase of enterprise AI will not be won by the companies that buy the most tools or drive the highest usage numbers," she said. "It will be won by the companies that make AI part of how work actually gets done, grounded in the right context, measured against real outcomes, and governed in a way that helps employees move faster without lowering the bar for quality."