The Work AI Index surveyed 6,000 knowledge workers: 3,000 in the United States and 1,500 each in the United Kingdom and Australia.
“So, it's 6,000 knowledge workers, 3,000 in the US. and then, uh, 1.5 K in each of the UK and Australia.”
Person
Rebecca Hinds
1 episodes · 15 quoted claims
Hinds says the Work AI Index survey was fielded in December 2025 and January 2026.
“We fielded it December and and January of 2025, 2026.”
The survey was developed with eight founding members of the Work AI Institute, which is Glean's internal research center.
“it was developed in partnership with our eight founding members of the Work AI Institute, which is our internal research center at Glean.”
Glean is a work AI platform that began as enterprise AI search, helping employees find relevant information within an organization.
“Sure, so Glean is a a work AI platform. Um, historically, we started in enterprise AI search, you know, solving the problem of how do we find the right, relevant information within an organization?”
Hinds says about 36% of AI sessions fail, meaning the worker must start over or do significant rework.
“We're seeing about 36% of all AI sessions fail, meaning a worker goes to use the technology and it's not successful.”
Hinds says roughly 40 to 41 percent of employees report shipping AI work they could not explain if asked.
“We see forty to forty-one of employees saying they ship AI work that they couldn't explain if asked.”
Hinds cites a Stanford study finding that 41% of Y Combinator AI startups are automating tasks people would prefer to keep human.
“There was a fascinating study that came out of Stanford a while back that found 41% of Y Combinator AI startups are automating things that people would prefer to keep human.”
Hinds says Rob Cross's research found high-performing organizations are up to 5.5 times more likely than less effective ones to measure and reward effective collaboration.
“high performing organizations, they're up to 5.5 times more likely than lower effective ones, less effective ones in measuring and rewarding effective collaboration.”
Hinds describes some organizations running hackathons that award prizes for prompt improvement, co-creation and peer feedback, not only for business impact.
“They also give prizes for improvement, best before and after prompt. They give prizes for co creation, feedback, peer feedback.”
Hinds says Paul Leonardi's research shows that when an AI assistant makes a mistake, people still blame the human user.
“when a human screws up, if I have an assistant and the assistant makes a mistake, a human assistant, we blame the human assistant. When an AI makes a mistake, we blame the human as well.”
Hinds says Nvidia CEO Jensen Huang does not hold one-on-one meetings with his direct reports, a practice she says works at Nvidia because of its strong mission clarity.
“The fact that Jensen doesn't have one on one meetings with his direct reports, that does not work in most organizations.”
Hinds says a CHRO at one of the largest healthcare organizations used AI to map all roles and found 70% of tasks overlapped with other roles.
“they've used AI to task map all their roles and just determine that 70% of all tasks are overlapping with other roles across roles.”
Hinds says she and Bob Sutton are writing a piece, hoped for Harvard Business Review, on how to flatten organizations well.
“Bob Sutton, and I were working on a piece right now. Hopefully it'll be in in Harvard Business Review.”
Hinds says the Work AI Index is planned as a pulse survey repeated roughly every six months to track changes over time.
“Every six months or so, the goal is to be able to track this longitudinally.”
Hinds says that when asked about work fully automated with AI, respondents averaged about 11 hours of weekly time savings, with other questions yielding 10 to 14 hours.
“On average, what we're seeing is people are reporting that based on the work output, they have fully automated with AI. It's about eleven hours of of time savings.”