AI power – Microsoft’s Work Trend Index finds AI “power users” generate and apply work differently, balancing automation with human judgment.
AI is no longer just boosting productivity at work; it’s creating a noticeable gap between people who use it deeply and those who don’t, according to Microsoft.
In its 2026 Work Trend Index report, Misryoum reports that Microsoft surveyed 20,000 knowledge workers who use AI at work.. The findings show that many AI users say the technology helps them spend more time on higher-value tasks. and a substantial share say their output has expanded compared with a year earlier.. The pattern becomes even clearer for a group Microsoft labels “frontier professionals,” suggesting that the benefits are not evenly distributed.
Insight: When AI is treated as a tool for capability-building rather than routine automation, workers tend to shift toward higher-stakes thinking and evaluation. That difference can compound over time.
Microsoft describes the dynamic as “capability add,” rather than simple efficiency.. In practical terms. the report highlights use cases such as helping teams identify and address software security vulnerabilities. enabling sales staff to ramp up faster ahead of customer meetings. and supporting work that previously would have been either too slow or too difficult to execute reliably.
Importantly, the report also indicates that top performers do not rely on AI blindly.. Misryoum notes that Microsoft says many frontier professionals sometimes complete tasks without AI to maintain their skills. and a meaningful portion of AI users deliberately decide—before starting work—what parts should be handled by AI and what should stay with human effort.
Insight: This “guardrailed” approach can help explain why power users outperform. Treating AI output as a draft and keeping humans in charge of quality control reduces the risk of errors becoming accepted work.
The report frames a broader shift in the kind of work humans do. with less time spent on step-by-step execution and more emphasis on setting direction. defining standards. and evaluating outcomes.. Even when no one directly supervises other employees. workers still end up practicing a supervisory role over AI systems—designing tasks for the tool. checking the results. and refining the workflow.
In that model, the responsibilities also extend into technology and security.. Microsoft points to the role of IT as a kind of control layer for “agent operations. ” emphasizing permissions and environments so that scale does not come at the cost of visibility.. Misryoum notes this builds on existing work in managing access and systems for people and applications.
Insight: As organizations widen AI usage, the teams that set rules, permissions, and evaluation practices may become just as critical as the teams using the tools day-to-day.
Still, Microsoft’s outlook acknowledges that AI will change jobs unevenly.. Misryoum notes the report stops short of promising uniform gains. pointing instead to the likelihood that some roles evolve. some decline. and new work emerges.. For employers, that complicates how AI expectations are set for employees.
Misryoum highlights that Microsoft discourages rigid performance checklists tied to how frequently someone uses AI. Instead, the company argues that appropriate adoption levels should follow employees’ expertise, experimentation, and the specific value AI can deliver in their workflows.
Insight: The key question for businesses may not be “how often AI is used,” but whether teams build the judgment, processes, and governance that let AI raise the quality of decisions and outputs.