Artificial intelligence is no longer just a productivity enhancer inside the workplace. According to a new global study from Boston Consulting Group, it is actively changing how jobs are structured, how employees spend their time, and what leaders now expect from their teams. The report suggests that many organisations are still treating AI as a tool rollout, even though the technology is already altering the mechanics of work itself.
BCG’s latest AI at Work report is based on a survey of 11,749 workers across 14 markets and multiple industries. The findings show that 72% of respondents believe AI has already changed the skills required in their roles, while nearly half say they now spend more time reviewing and directing AI-generated output than doing the work manually. That shift marks a clear departure from early AI adoption, when the focus was mostly on saving time at the individual level.
Work Is Becoming More Supervisory
One of the report’s strongest signals is that the role of employees is becoming more supervisory. As AI takes over simpler tasks, workers are being left with more complex assignments, more review responsibility, and a higher bar for what counts as good work. BCG found that 67% of respondents said AI has automated routine tasks, while 60% said expectations around acceptable output have already risen.
This is creating a workplace paradox. On one hand, 67% of regular AI users said the technology has improved job satisfaction. On the other hand, 41% reported greater cognitive strain and mental load. In practice, that means AI is reducing some forms of effort while increasing the pressure to check, interpret and manage what the systems produce.
Frontline Adoption Accelerates
The report also shows that AI adoption is spreading quickly among frontline white-collar workers. BCG found that 74% of these employees are now regular users, a sharp rise from two years ago. Emerging markets are leading this shift, with India, the Middle East, Brazil and South Africa recording stronger regular usage among frontline workers than countries such as the United States, France and Italy.
A big reason for this acceleration is that adoption is no longer limited to digital-native teams. Older workers, operational roles and employees in markets that were slower to digitise are now becoming regular users, which is helping AI move from pilot projects into day-to-day business activity. But the report also notes that companies are not always giving employees the structure needed to turn that time saved into better outcomes.
Time Saved, But Not Reinvested
BCG found that 42% of frontline employees say AI saves them at least one full workday each week. Yet 66% said they receive little or no guidance on how to use that freed-up time. More than half said the extra time is not being redirected toward strategic or higher-value work, which means the efficiency gains are not always translating into business value.
This is where many organisations appear to be falling short. The report argues that simply deploying AI tools is not enough if workflows, management systems and team responsibilities remain unchanged. Companies need to redesign work around AI, not just add AI on top of existing processes.
AI Agents Enter Daily Work
The study also points to the rapid rise of AI agents, which are beginning to move from experimentation into mainstream workplace use. BCG found that 30% of respondents already have AI agents integrated into workflows, up from 13% last year. Another 50% said their companies are actively piloting or experimenting with them.
That adoption is still uneven, though. Six in ten respondents believe AI agents could perform at least half their job within three years, but more than half say they still do not fully understand what AI agents are. Half also said their companies do not yet have clear oversight structures for human-AI collaboration, which raises questions around governance, accountability and decision-making.
Strategy Now Matters Most
BCG’s central conclusion is that strategic clarity matters more than tool adoption. The companies seeing the strongest results are not those with the most AI deployments, but those redesigning workflows end to end around how AI changes work. According to the study, organisations that pursued workflow redesign were far more likely to report business impact, higher employee satisfaction and greater productivity gains.
The report frames this as the “reshape/invent dividend,” where companies that rethink how work is organised capture more value from AI while also improving the employee experience. For CEOs, the message is direct: AI transformation now requires a clear strategy, measurable outcomes, employee involvement and governance that can evolve as the technology matures.
