A network storage location stores voice prints from a user of a wireless telephone. The voice prints are context-specific to likely responses to voice prompts from an interactive voice response (IVR) system. The voice prompts are communicated from the IVR to the wireless telephone. A voice selection is received from the wireless telephone. The voice selection is analyzed to determine a corresponding selection ID, and a voice print is retrieved from the storage location corresponding to the selection ID and to an ID of the wireless telephone.