Heidi Farris, CEO of ActivTrak, reveals data from the company’s Productivity Lab that challenges the prevailing narrative surrounding AI adoption. Tracking 120,620 employees across 1,009 organizations from Q4 2025 to Q2 2026, the research suggests that pursuing maximum AI maturity may not be the optimal goal for every business.
The Three Stages of Behavioral Maturity
The study categorizes users into three distinct stages based on how they actually interact with the technology:
- Stage 1 (Research Assistance): 27% of employees use AI as a search engine for answers and summarization.
- Stage 2 (Task Execution): 14% use AI to draft content, generate ideas, and handle routine tasks.
- Stage 3 (Workflow Integration): Only 2% have made AI an integral part of their daily professional workflows.
Overall, AI users remain a minority, accounting for 43% of the studied workforce.
The Productivity Paradox
According to the findings, "healthy utilization" peaks at 75% during Stage 2. Counterintuitively, once AI becomes fully embedded in workflows (Stage 3), healthy utilization drops by approximately 5 percentage points—reaching levels statistically indistinguishable from those who barely use AI at all.
Farris warns that pushing for deeper integration without clear guidance leads to two specific risks: runaway infrastructure costs and "operational disconnect." She highlights an internal example where employees used high-end models for simple email rewrites—the AI equivalent of hiring a superstar for a job that doesn't require one.
The Durability of AI Habits
A critical finding is the durability of adoption: 82% of employees who adopt AI continue to use it. Furthermore, once users move past casual use, they rarely return to traditional methods. This makes defining maturity targets a significant leadership responsibility, as these behaviors become locked-in over time. Farris emphasizes that the goal shouldn't just be increasing the 2% of power users, but coaching the 27% of novices toward task assistance fluency.
