Considering Costs of Interruption and Deferral in Routing Interpersonal Communications

We describe a prototype service that performs cost-benefit analysis about whether to route a phone call to a user versus reschedule the communication for later. The system balances the expected cost of interruption with the cost of deferral of a communication. We first present a decision- theoretic perspective on the handling of interpersonal communications. Then, we describe a prototype system named Bestcom-X that makes decisions about the routing of telephone calls coming into a corporate PBX, based on models of the cost of interruption that were built via machine learning. We then describe a derivative of Bestcom-X, named Bestcom-ET which is now being used actively used by approximately one thousand employees at the Microsoft Corporation.