Systems and methods for voice authentication are disclosed. In an embodiment, a computer system may determine that a user is eligible for establishing a voice authentication capability for a user account during a real-time audio communication between a user device corresponding to the user and a communication system associated with an electronic service provider. The computer system may enhance a recording quality of a portion of the real-time audio communication and record a voice sample for the portion of the real-time audio communication at the enhanced recording quality. The computer system may generate a voiceprint based on the voice sample and enable the voice authentication capability such that the user can be authenticated by voice in future audio communications with the communication system in a minimally intrusive fashion where normal conversation can be used to capture voice samples which can be compared to the voiceprint to authenticate the user.