Artificial intelligence is delivering real productivity gains for workers, but a significant chunk of that time is being quietly clawed back by a phenomenon researchers are calling "bot sitting" – and most organisations aren't even aware it's happening.
That's one of the headline findings from the Work AI Index UK 2026, produced by the Work AI Institute at Glean. The report found that while workers are saving an average of twelve hours per week through AI, around 6.3 of those hours are being spent on the manual labour required to make AI usable – feeding it context, supervising outputs, and cleaning up errors.
Speaking to UC Today, Dr. Rebecca Hinds, Head of the Work AI Institute at Glean, said the scale of bot sitting was the report's most surprising finding.
"We had a sense this was happening, but the sheer volume was surprising – and particularly in the UK, despite the UK being ahead in many areas, including governance."
The report also found that more than one in three AI sessions fail outright. Hinds attributes this largely to a context problem. "LLMs are trained on the vast corpus of the internet – they're often not trained deeply on our own work, our team's work, our organisation's work. That is the number one driver of bot sitting." She added that workers, fatigued by the constant effort of priming AI tools, are increasingly cutting corners – which only compounds the failure rate.
