Voice biometrics works by automatically comparing the voice of a customer on a call with a stored voice print of that customer to verify that they are indeed one and the same.
The idea being that this helps cut down on contact centre fraud whilst also improving customer experience by reducing call handle times and removing the need to ask security questions for identification purposes.
Voice prints are unique to each customer and when capturing them, the technology maps over 50 different physical and behavioural aspects of a voice to produce a unique and independent print. Physical characteristics takes into account the vocal tract, the shape and size of the mouth and nasal passage which generally don’t change over time. As for the behavioural factors this focuses on pronunciation, accent, pitch and intonation as well as speed of speech and the emphasis placed on words.
It takes approximately one minute of speech to create a voice print and there is no need for customers to actively enrol in any kind of programme, if you already have a recording of that customer you can simply create the voice print from that with no customer, agent or IT involvement.
Authentication is quick and will typically run in the background when a customer call comes in. Thresholds can be set to achieve caller authentication with scoring mechanisms set up as part of the calibration process. If the caller authentication does not match the required score, the usual security questions will dynamically pop up on the agents screen and you can revert to the usual tried and tested approach for verifying caller identity.
With the ability to build suspicious watch lists and create known fraudster voice prints you can be more proactive in your approach to combating contact centre fraud. For those contact centres operating in the financial and insurance sectors or particularly susceptible to fraudsters there is clearly a business case for investigating voice biometrics further.