Abstract: A system includes an interactive voice recognition (IVR) module arranged to perform a speech conversation with a first user and receive a first user identifier, where the speech conversation has an interaction context based on a subject matter of the speech conversation. The system includes a datastore arranged to store a group of active words associated with the interaction context and store first user voiceprints derived from pre-captured audio of the first user, where each active word is selected based on one or more selection criterion derived from conversations of a population of users. An automated speech recognition (ASR) module is arranged to perform speech recognition of the first user audio provided during the speech conversation. A voice biometric (VB) module is arranged to generate captured voiceprints and determine a similarity score based on comparisons of captured voiceprints with first user voiceprints.

Abstract: A system includes an interactive voice recognition (IVR) module arranged to perform a speech conversation with a first user and receive a first user identifier, where the speech conversation has an interaction context based on a subject matter of the speech conversation. The system includes a datastore arranged to store a group of active words associated with the interaction context and store first user voiceprints derived from pre-captured audio of the first user, where each active word is selected based on one or more selection criterion derived from conversations of a population of users. An automated speech recognition (ASR) module is arranged to perform speech recognition of the first user audio provided during the speech conversation. A voice biometric (VB) module is arranged to generate captured voiceprints and determine a similarity score based on comparisons of captured voiceprints with first user voiceprints.