A Glean survey of 6,000 workers in the U.S., U.K. and Australia finds that for every hour an employee spends getting a useful output from AI, they spend another hour making it usable: giving agents context, checking work, flagging mistakes, and cleaning up answers. Workers reported about 6.4 hours a week on this maintenance, more than a third (36%) of AI sessions fail outright and require a full restart or substantial rework, and 69% of AI users admit to shipping AI-generated work they had not reviewed, did not fully understand or could not defend if asked.
Why it matters to the HR leader: the productivity story most boards are told, that AI removes work, is on this data incomplete. AI is shifting work from visible production to invisible supervision, and that hidden shift lands hardest on the employees and managers least accounted for in the ROI math. Workforce plans built on assumed AI time-savings will overstate real capacity.
The signal underneath: Glean’s Rebecca Hinds, who heads its Work AI Institute, frames adoption as a “vanity metric” problem, where seats and prompts get measured but usable output does not. The companies that win, she argues, build human infrastructure around AI: training employees on when to use it and how to verify it, setting guardrails, and reinvesting reclaimed time into higher-quality work. In the company’s release she says the next phase of enterprise AI “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.” HR leaders should measure AI by outcomes per task.
Source: Glean.
Related on this site: Wells Fargo CEO Discusses the Complex Impact of AI on Employment and HR Strategies.
