ActivTrak: Moderate Use of AI May Be Optimal for Most Employees
Data from the ActivTrak Productivity Lab suggests that for most employees, the optimal approach may not be the most extensive implementation of artificial intelligence, but rather the regular use of AI for specific tasks. ActivTrak CEO Heidi Farris cited these findings in a column for Fortune.
The lab tracked 120,620 of the same employees across 1,009 organizations over three consecutive quarters—from the fourth quarter of 2025 to the second quarter of 2026. According to the data, 27% of participants used AI to search for answers and synthesize information. Another 14% used it to draft texts, generate ideas, and perform routine tasks, after which they reviewed and finalized the results. Only 2% integrated AI into their daily workflows. Overall, 43% of the surveyed employees used AI.
According to ActivTrak’s assessment, the “healthy usage” rate peaked at 75% at the stage when AI was regularly used to perform specific tasks. When the technology became an integral part of workflows, this metric dropped by approximately five percentage points—to a level that was not statistically different from that of employees who rarely used AI.
Read more breaking news on the UA.News Telegram channel.
Farris noted that deeper AI adoption does not always translate to better business outcomes. In her view, companies may be spending more on more powerful models, tokens, and infrastructure for tasks that do not require them. She also cautioned that employees may create complex AI workflows that speed up individual tasks but do not improve broader business processes.
As an example, ActivTrak cited rising costs associated with the use of Anthropic. After conducting an analysis, the company discovered that employees were regularly using the latest and most powerful model to draft emails to clients, even though such tasks did not require such a sophisticated tool. In response, the company developed internal guidelines for selecting the appropriate model based on the task at hand.
According to the lab’s data, 82% of employees who started using AI continued to do so. Farris believes that before a large-scale rollout, companies should analyze their existing workflows and test targeted changes, selecting tools based on employee roles and business needs.
