The growing adoption of Artificial Intelligence in companies has revealed a less discussed reality: AI systems continue to require human supervision, transforming work rather than eliminating it. It is in this context that the concept of botsitting emerges, defined as the time and effort dedicated by employees to monitoring, correcting, guiding, and supervising AI systems.
A recent study by Harvard Business Review, published in August 2026, concluded that workers spend an average of 6.4 hours per week on botsitting, equivalent to practically a full working day. This time represents 37% of the total time dedicated to interacting with AI, slightly above the 36% spent using the technology to produce work.
The article, authored by Margarida Duarte, Senior Consultant at EY in the People Consulting area, warns that this reality creates a new layer of invisible work that requires skills that technology cannot replace: critical thinking, analytical ability, contextual knowledge, judgment, and accountability. The question also arises at the leadership level, raising questions about responsibility when a system makes a mistake, how to measure the performance of professionals who supervise digital agents, and what skills should be developed.
EY works daily with organizations seeking to transform AI adoption into sustainable results, through the redesign of operating models, team preparation, definition of governance mechanisms, and management of the change necessary to integrate AI responsibly. The article concludes that the true potential of AI depends on the ability to combine technology, processes, and human talent responsibly and sustainably.
