Verify Clients in 3 Seconds Based on Their Voice

Verify Clients in 3 Seconds Based on Their Voice

February 25, 2020

By in Blog

How long does it take your contact center agents to verify clients over the phone? 30 seconds, a minute, or even longer? Stop wasting everyone’s time with a neverending list of security questions and authenticate callers in just 3 seconds based on their voice!

Year after year, Phonexia researchers push the accuracy of speaker recognition systems to its limits, and 2020 is no exception! We are glad to announce that our latest generation of Deep Neural Network (DNN)-powered Speaker Identification technology can verify speakers reliably in just 3 seconds of net speech based on voice biometrics.

This is a remarkable milestone in commercially available speaker verification systems. Shorter verification time means significantly reduced contact center costs and leads to greater customer experience without compromising security.

It opens an entirely new chapter of customer experience (CX) and account security for contact centers, conversational AI platforms, and other voice biometrics solutions powering digital transformation.

What took our previous speaker identification technology 10 seconds to identify, our latest speaker identification engine can recognize in just 3 seconds.

The chart below illustrates how the accuracy of Phonexia Speaker Identification technology has evolved over the years:

The Accuracy of Phonexia Speaker Identification Technology in 2015, 2019, and 2020.

Notice the mind-blowing difference between the accuracy of Phonexia Speaker Identification technology from 2015 (the world’s top-class technology back then), 2019, and 2020.

In just 3 seconds, the 2019 technology recognized speakers with the same accuracy as the 2015 technology achieved in 30 seconds, only to break the record again this year with the release of our most advanced voice biometrics technology ever, recognizing speakers after only 3 seconds of net speech and with the accuracy that the 2019 technology offered after 10 seconds.

The accuracy of Phonexia Speaker Identification technology can be improved further through adaptation using customer-provided data.

How Much Can You Say in Three Seconds?

As a worldwide provider of cutting-edge voice biometrics and speech recognition technologies, Phonexia is well aware of the increasing need for robust voice verification that would be able to identify calling clients with high accuracy based on extremely short net speech.

Imagine how much time you could save on both sides by skipping the “what was your mother’s maiden surname” type of security question. How much could you reduce contact center costs by being able to skip tens of seconds during each call’s authentication phase? What if you could recognize fraudsters calling to your contact centers, claiming to have some other identity? Would you like to improve customer experience and security at the same time?

Digital transformation demands state-of-the-art approaches when it comes to authentication over the phone, and frictionless voice biometric authentication is definitely one of them—it offers a perfect balance between security measures and a seamless customer experience.

Share now!

Recent posts