The many benefits that cell phones provide are at times overshadowed by the problems they create, as when one person’s cell phone disrupts a group activity, such as a class, meeting or movie. Cell phone interruption is only highlighted by the ever increasing number of mobile devices we carry. Many tools and techniques have been proposed in order to minimize interruption caused by mobile devices. In the current study, we use calendar information to infer users’ activity and to automatically configure cell phones accordingly. Our in-situ experiment uses PDAs that run a cell phone simulator to examine the feasibility and design factors of such a solution. Our results show that both structured activities and appropriate cell phone configuration can be predicted with high accuracy using the calendar information. The results also show consistent mapping of activities to configuration for each individual. However there was a poor consistency of mapping activity to configuration across different participants. We discuss the results in relation to inaccuracy, spontaneous activities, and user reactions.