Airflow and pollutant transport models are commonly based on the approximation that air is instantaneously well mixed in a zone. In many circumstances this approximation is unsatisfactory. We present a semi-empirical model to predict the time required by a pollutant to disperse in a room, owing to the effects of room occupancy (room mixing induced by thermal plumes, stirring by people walking about and breathing). We base our estimates on previous experimental work correlating the mixing tome for a pollutamt pulse released in a room to the mechanical energy supplied to the room air. Our results suggests that people moving about in a room can induce rapid mixing. For example, the mixing timeresulting from a person walking at 1.4 m/s in a 30 m^3 room is predicted to be about 4 minutes . In this case, the assumption that air is well mixed could be reasonable.