37% of AI time savings eroded by need to review and revise outputs

While AI has increased gross productivity, net gains have been eroded by the time and effort needed to review and correct AI outputs. 

Processing Content

According to a report from finance and HR platform Workday, roughly 37% of the time saved through AI is being offset by time spent correcting, clarifying or rewriting low-quality outputs. For every 10 hours of efficiency gained through AI, nearly four hours are lost to rework. For highly engaged employees who frequently use AI, Workday believes this translates into about 1.5 weeks per year devoted to correcting bad AI outputs. And so while 77% of employees say they are more productive due to AI use over the past 12 months, and 85% of employees personally save between one and seven hours per week on their tasks, this does not necessarily translate into better overall performance. 

“This hidden loss highlights a critical blind spot in how organizations assess AI performance. Most leaders focus on gross efficiency—how much time AI saves. But this metric alone obscures the real picture. When time lost to rework is taken into account, the net value of AI is often much lower than expected. To capture AI’s real return, organizations must move beyond measuring hours saved and begin accounting for outcomes achieved. Net value—time saved minus time lost—provides a more accurate view of whether AI is improving how work gets done, or simply accelerating activity without improving results,” said the report. 

One factor behind this paradox, according to the report, is a lack of training. While 66% of leaders cite skills training as a top priority, only 37% of employees facing the most rework say they’re getting it. Another factor is that many roles have not been updated to include AI-related skills, which the report described as “forcing employees to use 2025 tools within 2015 job structures,” something that often increases effort, rather than reducing it. 

The report highlights an issue that has grown apace with AI adoption, namely what has been called low-quality AI “workslop.” It is generally defined as AI-generated work content that mimics good work, but lacks the substance to meaningfully advance a given task, such as reports that look polished and read well but make no sense. Of 1,150 U.S.-based full-time employees across industries polled, 40% report having received such content in the last month. Typically, those who receive it have to then spend time verifying information, correcting errors and otherwise doing work that the person who sent it should have already done.

Employees said such low-effort, low-quality content makes up about 15.4% of all content they receive at work. The researchers found that people spend an average of nearly two hours dealing with each instance of workslop. Based on participants’ estimates of time spent, as well as on their self-reported salary, the researchers found these incidents carry an invisible tax of $186 per month per person. 

Organizations need to measure productivity by outcomes, not just hours saved; quality, accuracy, and decision speed must be considered alongside time savings, according to the report. They also should redesign roles for AI versus just bolting AI onto job structures, which may involve clarifying when AI assists, when humans decide, and how success is measured. Finally, the report recommended that leaders reinvest whatever time is saved into its people. 

“Organizations that realize net gains from AI explicitly authorize employees to use saved time for activities that improve collaboration, learning and strategic thinking—not just increased task volume,” said the report. “Companies should train managers to recognize high-friction points and authorize employees to reinvest time in connection and problem-solving.”