Workday study: AI productivity gains fizzle out due to reworking
A new global Workday study shows that around 40 percent of the time saved with AI is lost due to reworking inferior results. Successful companies invest the time they save in training and developing their employees. In Switzerland, 44% of employees use AI tools, but many spend up to two hours a week on corrections.

Workday publishes a global study that reveals a key paradox of AI usage: while artificial intelligence delivers significant time savings, many organizations are not fully capturing its value. Much of the time saved is eaten up by rework - correcting errors, rewriting content and validating the outputs of generic tools.
The AI productivity paradox
The study «Beyond productivity: What AI is really worth» shows clear discrepancies between the promise of productivity and the actual benefits. In Switzerland, around 71% of employees report that they have been significantly more productive in the last twelve months thanks to the use of AI. Overall, more than a third (44%) use AI tools or technologies in their daily work.
But the flip side: around 54% say that they spend one to two hours a week on reworking the time saved with AI. This leads to a false impression of productivity and ROI. AI contributes by increasing capacity - but all too often roles, skills and processes have not evolved to turn that capacity into consistently better results.
Investment gap in further training
Another problem can be seen in training measures: Around 63 percent of managers surveyed say their organization has prioritized reinvesting AI-powered efficiency gains in skills training and education. However, only 36 percent of employees report increased organizational investment in new or expanded training and development programs. This shows a clear disconnect between leadership intent and employee experience.
Although most organizations agree that the productivity gains from AI should benefit employees, in practice these investments often flow elsewhere. Companies are more likely to reinvest savings from AI in technology (39 percent) than in employee development (30 percent). Instead of using the time gained specifically for skills development, the workload is often simply increased (32%).
What successful companies do differently
The organizations that actually achieve measurable success take a different approach. Employees with positive AI experiences are significantly more likely to use the time saved to increase the value of their work - for example, through deeper analysis, more informed decisions and strategic thinking (57%), rather than simply taking on additional tasks. They are also significantly more likely to have benefited from extended qualification and further training measures (79%).
«Too many AI tools push the hard questions of trust, accuracy and repeatability onto individual users,» says Gerrit Kazmaier, President of Product and Technology at Workday. «At Workday, we've spent years delivering AI as simple, human-centered solutions - not pure technology - so customers don't have to wire everything up themselves and check every answer. Our philosophy is that AI should do the complex work in the background so humans can focus on judgment, creativity and connection. That's how organizations turn AI-powered speed into sustainable, human-led advantage.»
The study makes it clear that reinvesting in people is the fastest way to reduce rework, improve results and turn AI speed into sustainable business value. The organizations that realize the most value from AI treat saved time as a strategic resource and invest in upskilling their teams.
About the study
The data comes from the global study «Beyond Productivity: What AI is Really Worth», a survey conducted by Workday and carried out by Hanover Research in November 2025. The study included 3,200 respondents from North America, Asia Pacific and Europe, the Middle East and Africa. All respondents were full-time employees in organizations with annual revenues of $100 million or more and were active users of AI technology.
Source: Workday






