AI adoption inside companies has stopped being the hard part. Making that adoption pay off is where organizations are stalling, according to Glean’s Work AI Index 2026, a global study built with researchers from Emory, Notre Dame, UC Santa Barbara, UC Berkeley, University College London, UNC Charlotte, and Stanford. The study finds that 87 percent of digital workers now use AI at work and 75 percent say it makes them more productive, saving roughly 11 hours a week through automation. Yet only 13 percent say their organization is performing significantly better as a result.
Glean’s researchers trace that gap to a form of labor most HR systems do not track. They call it “botsitting”: the work of feeding AI missing context, checking its outputs, debugging mistakes, rerunning prompts, and cleaning up confidently wrong answers. Workers now spend an average of 6.4 hours a week botsitting, nearly a full working day. When that time goes unbudgeted and unrewarded, the study finds workers start cutting corners, a behavior Glean labels “botshitting”: shipping AI-generated work they have not reviewed and could not defend if asked. Today, 69 percent of AI users admit to doing it.
The insight HR leaders should take from this is not that employees are gaming the system. It is that the workers getting the most value from AI are also the ones spending the most time managing it. Glean’s data show high achievers, people who report both productivity and quality gains, spend more time botsitting than low achievers (40 percent versus 33 percent) and are 18 percent more likely to deliberately hold back from using AI on certain tasks. That is a judgment skill, invisible to most performance dashboards. As earlier research on AI babysitting time has already shown, those hidden hours are becoming a workforce planning problem HR cannot keep leaving off the books.
Source: Glean Work AI Institute
