Employees save up to seven hours weekly with artificial intelligence (AI), but organizations are squandering productivity gains on fixing errors rather than strategic work, according to new research from Workday Inc.
The study of 3,200 AI users in North America, Europe, and Asia-Pacific reveals a troubling “productivity paradox.” While 85% of employees report saving between one and seven hours weekly using AI tools, nearly 40% of that value evaporates through rework and misalignment. Only 14% of workers consistently achieve clear, positive outcomes from AI use.
The heaviest AI users face the steepest challenges. Nine in 10 daily users overwhelmingly believe the technology will help them succeed, yet 77% say they must review AI-generated work as carefully as human output, or more so.
“Organizations are capturing speed but losing outcomes,” the report concluding, highlighting a fundamental disconnect between AI adoption and business value. Rather than freeing employees for high-impact work, reclaimed time increasingly goes toward correcting poor-quality AI output and reconciling conflicting guidance.
The problem stems partly from outdated workplace structures. Less than half of job roles have been updated to reflect AI capabilities, forcing employees to use 2025 technology within 2015 job descriptions. Workers struggle to reconcile faster output with unchanged expectations around accuracy, judgment, and risk management.
A training gap compounds the issue. While two-thirds of leaders cite skills development as a top priority, only 37% of employees experiencing the most rework report receiving adequate training, exposing a stark divide between leadership intentions and employee reality.
Perhaps most troubling is how organizations reinvest AI gains.
Despite four in five leaders acknowledging that workforce development is crucial, companies allocate a larger share of AI cost savings to technology (39%) than to people (30%). Time savings are more frequently absorbed by increased workloads (31%) rather than employee development and upskilling (26%).
“Real AI ROI comes from investing in the people — skills, roles, and decision-making — as much as the technology,” researchers said.
The study recommends three corrective actions: measuring productivity by outcomes rather than hours saved; redesigning roles specifically for AI collaboration instead of bolting technology onto existing structures; and explicitly authorizing employees to use saved time for learning and strategic thinking rather than simply adding more tasks.
