AUTOMATICALLY MONITORING FOR VOICE INPUT BASED ON CONTEXT

In one implementation, a computer-implemented method includes detecting a context associated with a mobile computing device; determining, based at least on the context associated with the mobile computing device, to automatically switch from a first mode of operation in which ambient sounds are not monitored to determine whether the ambient sounds include one or more trigger terms that correspond to a request from a user for the mobile computing device to perform an operation, to a second mode of operation in which the ambient sounds are monitored to determine whether the ambient sounds include the one or more trigger terms that correspond to the request from the user for the mobile computing device to perform an operation; in response to determining to automatically switch to the second mode of operation, activating one or more microphones to receive the ambient sounds, and activating a speech analysis system for determining whether the ambient sounds include the one or more trigger terms that correspond to the request from the user for the mobile computing device to perform the operation; receiving data indicating that the speech analysis system has determined that the ambient sounds include the one or more trigger terms that correspond to the request from the user for the mobile computing device to perform the operation; and based at least on receiving the data indicating that the speech analysis system has determined that the ambient sounds include the one or more trigger terms that correspond to the request from the user for the mobile computing device to perform the operation, providing an output that is responsive to the request.