Indentico and its Technology Partners deliver the Indentico™ Speaker Identity System and services anywhere in the world and in any language.Our Speaker Identity System is a voice driven security solution for voice biometrics systems integrators featuring:
Impostor Rejection Technology
How Voice Biometrics work
Conceptually, voice biometrics is very simple. First, an individual enrolls their voice in the system, typically by repeating an account number and/or date of birth, three times.
From this, the system creates a “digital model” of that individual’s voice, often referred to as “a voiceprint”. The voiceprint is not a voice recording; but a digital representation of the way that individual’s voice sounds.
Then on a subsequent call, the caller makes a claim to an identity, most often by quoting their account number. The voice sample spoken by the caller is compared against the voiceprint associated with that account number.
If it is the same individual and there is a good match between the voice and voiceprint, the likelihood score computed will be high, resulting in the system accepting the caller’s claim to identity.
However, if it is a fraudster, using stolen account numbers for example, the voice match will be weak, resulting in the claimed identity being rejected by the system. At this point the caller can be transferred to a special section of the bank trained to handle failed authentication attempts.