AI is reshaping day-to-day work inside organizations, with new Ipsos survey data in collaboration with research firm Epoch AI, showing one in five full-time workers say AI has replaced parts of their role.
“When we actually look at what people report for their AI usage, we do see augmentation and automation effects,”
Caroline Falkman Olsson, Resarch Lead, said.
The findings suggest AI adoption is meaningfully changing work to automate tasks. However, while usage continues to expand rapidly, the same data points to a growing question for enterprises: whether these efficiency gains are translating into measurable business value at scale.
The survey points to efficiencies in the tasks it is used for, but those tasks remain light or surface level.
Employees Are Already Changing the Way Work Gets Done
The survey, which polled more than 2,000 adults, found that 20% of full-time workers reported AI had replaced elements of their job. At the same time, 15% said AI had created new forms of work for them, suggesting automation is currently outpacing the creation of new tasks or responsibilities.
Researchers behind the study said the findings show both augmentation and automation are already taking place simultaneously inside workplaces. Rather than replacing entire positions outright, AI appears to be steadily absorbing smaller components of jobs, particularly routine administrative activities and repetitive knowledge tasks.
The data also revealed that much of this adoption is happening outside formal enterprise governance structures. Around half of employees using AI for work said they relied on personal subscriptions or free public tools instead of employer-approved platforms, raising concerns around oversight, compliance, and visibility into how AI is being integrated into workflows.
At the same time, organizations that provided employees with company-funded AI tools saw significantly higher levels of regular usage. The findings suggest access and enablement remain major factors in determining how deeply AI becomes embedded into everyday operations, particularly as workers increasingly seek ways to streamline repetitive work independently.
AI Is Changing Workflows, but Productivity Gains Remain Hard to Prove
While the survey points to meaningful workplace disruption, it also suggests AI is still being used in relatively narrow and low-impact ways rather than reshaping higher-value work.
Most employees reported using AI for short, routine activities such as summarizing content, drafting communications, or generating ideas, often completing only one or two AI-assisted tasks even on their busiest day. Only a small proportion said they use AI heavily, indicating that adoption is still concentrated around lightweight, repetitive functions rather than deeper workflow transformation.
This matters because it suggests AI is not yet being consistently applied to the areas where it could have the greatest organizational impact, such as decision-making support, customer strategy, complex analysis, or process redesign. Instead, it is primarily being used to accelerate administrative and convenience tasks that sit at the edges of core business value.
However, the direction of travel is still significant. The same data shows that AI is already replacing elements of work faster than it is creating new types of tasks or responsibilities. Even if current usage is concentrated at the lower end of complexity, that imbalance suggests the gap between automation and task creation could widen as adoption deepens and organizations begin applying AI to more critical workflows.
In other words, AI is increasingly automating work, but currently at the margins of value creation. The key question is whether that limited use represents an early stage of adoption, or the beginning of a broader shift in which automation begins to extend into higher-value work as capabilities and trust in the technology grow.
Enterprises May Need a New Framework for Measuring AI Success
While organizations are investing heavily in AI tools, the challenge is increasingly shifting from deployment to understanding how deeper integration changes the structure of work itself.
For many businesses, the immediate gains from AI are still concentrated in small efficiency improvements. Employees are using the technology to complete routine tasks faster, but those gains are often incremental rather than transformative at an organizational level. As a result, productivity improvements can appear modest even as usage continues to rise.
However, the trajectory suggested by the data points to something more significant over time. If AI becomes more deeply embedded into core workflows, the balance between tasks automated and new tasks created could shift more decisively towards automation.
Organizations that provide employees with company-funded AI tools see significantly higher levels of regular usage, indicating that access and enablement are key determinants of adoption depth.
The implication is that deeper automation gains will depend not just on capability, but on whether enterprises actively enable structured access to AI at scale. Without that, usage may remain uneven and confined to lighter tasks, limiting the extent to which AI can meaningfully reshape workflows.
