Dive Brief:
- Although three-quarters of knowledge workers say AI makes them more productive, only 13% report that the technology has significantly improved their company’s performance, a survey from Glean’s Work AI Institute found. Glean surveyed 6,000 digital workers and used insight from AI leaders who use its platform for its inaugural Work AI Index report.
- On average, automation saved workers about 11 hours a week, the survey found, but employees reported spending much of that saved time managing AI. Of hours spent interacting with AI, workers said they spent slightly more time managing the tools than using them to produce work. Only 27% of time was spent learning how to use the tools and building agents.
- Successful enterprise AI adoption hinges on the human infrastructure around it, the report found. To improve use cases, leaders must work to ground AI in an enterprise context, train employees on use cases, treat shadow AI as a signal that company-approved tools are falling short, and build governance into daily decisions.
Dive Insight:
Alhough AI is taking formerly human-powered tasks off of workers’ plates, employees are now spending time doing low-visibility tasks to make AI outputs usable, such as giving agents context, checking their work, flagging mistakes and cleaning up answers.
Workers spent nearly six and a half hours per week on these maintenance tasks, the report found. And monotonous work can easily turn into mistakes. If workers stop carefully reviewing outputs or verifying that AI’s recommendations make sense — as 69% reported they do — mistakes will slip through the cracks.
“Too many companies are treating AI adoption like a vanity metric — more seats, more prompts, more usage,” said Rebecca Hinds, head of the Work AI Institute at Glean, in the report.
But more AI usage doesn’t equal productivity or tech transformation, she said. The report found that for every hour an employee spends getting a useful output from AI, they spend another hour making it usable. More than a third of AI sessions fail completely, requiring employees to start over or largely rework the tasks, Glean found.
If employees are spending too much time on AI management, companies haven’t eliminated work, they’ve just created new kinds of work and more overhead for employees and managers, Hinds said.
Enterprises that make AI a part of how work actually gets done will be more successful than ones that use AI just for the sake of it, Hinds said. Successful companies build more human infrastructure around their AI use, training their employees on how and when to use AI and the guardrails around it.
They also learn to reinvest time saved from AI into higher-quality, human-centered work and stronger AI skills, instead of using the most AI possible, the report said.
Success for enterprise companies is more likely with foundations at the individual, team and organization level.
“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,” Hinds said.
