AI is stretching the workday. Four years after AI became broadly available as a commercial product, productivity returns have mostly failed to materialize, and workers forced to adopt these tools are becoming less efficient as time goes on.
An analysis from ActivTrak, a workplace surveillance company, found that white-collar workers who use AI put in up to 90 minutes more work each day than non-users. Analysts examined anonymized work activity from over 9,100 employees at over 450 companies between April and September.
Non-AI users recorded an average active workday of 7.6 hours. Heavy AI users recorded between 8.4 and 9.1 hours, suggesting that AI adoption can add billable time instead of removing it — a curious outcome for technology sold as a shortcut.
The analysis also linked heavy AI use to more interrupted work than light or non-users. AI may help employees complete individual tasks faster, but it also brings more application switching and shorter focus sessions, according to Gabriela Mauch, head of ActivTrak’s productivity research division.
Individual gains are not becoming team gains
Miro research points to the same gap from another direction. Its survey of over 9,500 leaders and workers found that 83% believe AI has improved their individual productivity, while 73% say the speed of output has increased.
Quality tells a less flattering story. Only 56% said AI improved the quality of output, and 55% reported an increase in “AI slop” requiring additional human review and rework. Nearly 1 in 10 said decision-making had become worse because of increased AI usage.
A majority said AI had not improved decision-making, at 52%, or collaboration between colleagues and teams, at 60%. The survey also found that 69% of AI-assisted work happens in private sessions rather than shared spaces, turning a supposedly collective transformation into a collection of individual experiments.
The organizational barriers are familiar but now have machine-generated output attached to them. Forty-nine percent of enterprises struggle with information silos, employees cannot access shared context across systems, and 40% of leaders rate visibility into AI use as poor or non-existent.
As many as 58% report compliance and security concerns without unified governance. Meanwhile, 44% of workers report more context switching, 59% face a higher volume of AI-generated output to review, and 60% worry that role boundaries will blur; 47% fear related roles will merge.
The human cost reaches beyond workload. Forty-four percent report concern about declining collaboration culture, and 1 in 5 say work feels lonely because of increased AI usage. Faster output is easy to measure; the damage from fragmented attention is less cooperative.
Efficiency can become an innovation trap
Dave Ross, Product Evangelist Leader at Miro, said faster execution is a benefit, but decision-making and decision quality are the real bottleneck. He described a need for places where AI and people can collaborate on critical decisions instead of working through fragmented, disjointed single-player tools.
Gartner refers to these solutions as AI workhubs, and two-thirds of enterprise decision-makers plan to implement one within 12 months. The goal is to give teams and their agents shared context, rather than adding another private tool to an already crowded software stack.
The efficiency push carries a deeper risk. Companies that use AI mainly to eliminate variation and maximize efficiency can undermine the experimentation and creativity that drive innovation — the same productivity dilemma William Abernathy of Harvard Business School identified when increased efficiency came with decreased innovation.
Process management programs such as Six Sigma reduce variance in company routines. Management scholars Mary Benner and Michael Tushman documented the tradeoff: the more process management a company used, the more its innovation relied on what it already knew instead of creating new approaches and drawing on new ideas.
Workday has told investors that customers should pay for AI agents out of labor and services spending, not just software budgets. Its CEO also said Workday was looking to replace new hires with AI, making the efficiency argument unusually direct.
The evidence so far does not show a clean replacement of labor with machine productivity. It shows longer active workdays, more output to review, more switching between applications, weaker collaboration, and pressure on decision quality. AI can make work faster — but without shared context and room for experimentation, it may only help organizations produce more noise before closing time.
Based on
- Employees Who Use AI End Up Working Vastly More Time Per Day — futurism.com
- Bosses loved AI at work — until they drowned in slop — africa.businessinsider.com
- Individual AI Productivity Soars but Organizational Impact Stalls | AFP.com — origin-zh9p7m-www.afp.com
- Too much AI can stifle innovation – The Japan Times — japantimes.co.jp
