UAE Staff Save 3.6 Hours a Week With AI, Then Spend 3.4 Fixing It
Zeta42 · 2 October 2026
A Workday report finds UAE employees save 3.6 hours a week using AI but spend 3.4 hours correcting its output. The gap between the two is a skills problem.
A gain of about 12 minutes
UAE employees save an average of 3.6 hours a week by using AI, but spend an average of 3.4 hours clarifying, correcting or rewriting low-quality AI output, according to a Workday report covered by Khaleej Times. Taken at face value, the two averages leave a net gain of about 12 minutes a week.
The report, "Beyond Productivity: Measuring the Real Value of AI in the UAE", looks at how organisations measure AI's impact beyond the time it saves. Adoption is not the problem. Workday found that 90 per cent of UAE employees use AI at least once a week; 58 per cent use it several times a week and 18 per cent every working day. More than half of directors and C-suite executives use it weekly. Just 0.2 per cent of respondents said AI had failed to produce meaningful time or cost savings.
So nearly everyone uses AI and almost nobody says it has failed to save them time or money, yet on these figures most of the time it saves goes straight back into checking and repairing its work. Khaleej Times drew the obvious conclusion: the next challenge for employers may be less about getting people to adopt AI and more about reducing the time needed to check and correct what it produces.
Why the fix time is a skills problem
Much of the rework on AI output starts before the model writes a word. A vague request, a missing file, no agreed idea of what a good answer looks like, a task that was never suited to AI: each produces a draft someone has to rewrite, and each is a habit that training can correct.
Some checking should stay. Zakaria Haltout, Workday's Group Vice President for the Middle East, Turkey and Africa, said the figures show UAE organisations are moving "from AI experimentation to everyday use". "But lasting value depends on trust, human oversight, and continued investment in skills," he said. Oversight is worth keeping. What tends to eat the hours is oversight without a method, where a person rereads every line because they have no idea where the model tends to go wrong.
Good training works on both ends of that. People learn to brief a model with the context and files it needs, so the first draft is closer to usable. They also learn where to look when they check, so review becomes targeted instead of a full rewrite. Both are learned on real work rather than demos, which is why Zeta42's department cohorts run for four to six weeks around the team's own tools and files.
Where the saved hours go
Workday also asked what organisations do with the efficiencies AI creates. Around 33 per cent use them to handle bigger workloads without adding headcount. Another 32 per cent direct the time and money saved towards training, upskilling and reskilling employees, and 32 per cent invest in workload management, workplace flexibility and employee well-being. Only 24 per cent said the savings mainly go to cutting costs.
The organisations putting savings back into training have the right instinct, and the 3.4 hours show where to aim it. Training that stops at how to use a tool raises adoption, which in the UAE is already high. Training that ends in a changed process is what should bring the fix time down. Zeta42's "AI at Work" track ends with a working change to a real process for that reason.
The pressure to show results is real. Angelique De Vries-Schipperijn, Workday's President for EMEA, said UAE organisations are pursuing transformation while needing to demonstrate results and maintain trust and control, particularly in large and regulated businesses. The research was released as Workday launched operations in the UAE, including a new Dubai office. Twelve minutes a week is a hard result to present to a board. The hours lost to fixing are where the real return is waiting.
Source: khaleejtimes.com