AI ‘productivity paradox’ as rework erodes time gains

Workday has published new global research that suggests organisations lose a significant share of AI-related time savings to rework, with employees spending hours correcting, clarifying and rewriting low-quality output.

The report argues that many employers have not updated job structures and skills programmes to match the pace of AI adoption. It says this reduces the business impact of faster task completion.

Workday based the findings on a survey of 3,200 full-time employees at organisations with annual revenue of USD $100 million or more. Respondents were active users of AI technology across North America, Asia-Pacific and Europe, the Middle East and Africa.

Productivity trade-off

The research frames the issue as a "productivity paradox". Employees report that AI tools save time each week. The report says that time does not consistently translate into improved outcomes.

Workday reported that for every 10 hours of productivity gained through employee adoption of AI, organisations "pay back" four hours correcting and rewriting low-quality output. The report describes the rework as checking accuracy, fixing mistakes and rewriting generic AI-generated content.

The survey found that 85% of employees say they save one to seven hours per week using AI. At the same time, Workday said nearly 40% of AI time savings get lost to rework.

Only 14% of employees "consistently get clear, positive net outcomes from AI," according to the report.

Workday said frequent users feel a greater burden from review and verification. It reported that employees who use AI every day remain optimistic about its impact on their success. It also said that 77% of daily AI users review AI-generated work just as carefully as human-created work, or more carefully.

The report also highlighted differences by age group. Workday said employees aged 25-34 account for 46% of those dealing with the most AI rework.

Jobs and structures

The research points to organisational design as a core issue. Workday said AI increases capacity. It said roles, skills and processes have not evolved enough to convert that capacity into better results.

The report said that in most organisations, fewer than half of job roles have been updated to reflect AI capabilities. It put that figure at 89% of organisations. Workday described this as employees "using 2025 tools inside 2015 job structures".

The findings also raised questions about where organisations invest the time saved through AI adoption. Workday said companies were more likely to reinvest AI savings into technology than employee development. The report said 39% put AI savings back into technology, compared with 30% into employee development. It also said 32% increased workload rather than using time saved for training and development.

Training gap

The report describes a disconnect between leadership intent and employee experience on training. It said 66% of leaders cite skills training as a top priority. It also said only 37% of employees experiencing the highest amount of rework report access to that training.

The research also compares the behaviour of employees who report positive AI outcomes with those who do not. Workday said employees with positive AI outcomes were more likely to use time saved to increase the value of their work. It said these employees focused on deeper analysis, stronger decision-making and strategic thinking. It put that figure at 57%.

Workday also said employees with positive AI outcomes were more likely to have had increased skills training. It reported that figure at 79%.

UK picture

Workday also published UK findings from the research. It said three-quarters of leaders in the UK reported that their organisation had prioritised reinvestment of AI-driven efficiency gains in skills training and development. It put that figure at 79%.

The report said one-third of employees in the UK use AI tools and technologies once a day or more. It reported that figure at 39%.

It also said that UK employees typically spend one to two hours per week clarifying, correcting or rewriting low-quality AI outputs. It reported that 45% of employees fall into that range.

Workday view

Workday's President, Product and Technology, Gerrit Kazmaier, said the company sees quality, trust and repeatability as central to realising value from AI in the workplace.

"Too many AI tools push the hard questions of trust, accuracy, and repeatability back onto individual users," said Gerrit Kazmaier, President, Product and Technology, Workday.

The report said organisations that see "real returns" treat time saved as a strategic resource and reinvest in skills and redesigned roles. It said the largest opportunity sits in improving how employees use AI in judgement-driven work.