How to: Reduce contact centre verification time by 80%
Verification of identity has changed a lot over the years. As the world has shifted away from face-to-face interactions to phone and digital solutions, so has the way in which we identify and verify who we are.
This has led to a problem – securely verifying a person’s identity takes time, and as we all know, time is money.
According to Microsoft’s 2019 survey on global customer service, 31% of customers cite having to repeat or provide their data multiple times as the most frustrating aspect of poor customer service.
This is often due to problems with automatic speech recognition (ASR) based systems, which are notoriously bad at recognising proper names. This means not only annoyance for end-users but also a need for call centre agent’s time in order to deal with issues arising from this, such as error recovery and rectifying customer satisfaction.
Security can’t be compromised for speed
As the use of high-level authentication is usually reserved for services such as financial companies, it is imperative that no corners can be cut in the hope of speeding up the process. In fact, with cybercrime spiking hugely throughout the pandemic, businesses need a solution that is MORE secure than traditional methods, while accelerating the process and creating a streamlined end-user customer experience.
We all lead busy lives, the last thing your end user wants is spending time wracking their brain to remember the make of their first car or what password is needed from their 100+ accounts.
Customers call contact centres wanting solutions. Voice biometrics allows contact centres to deliver these quickly and securely, removing barriers instead of creating them.
How does Aculab’s Voice Biometrics work?
Instead of requiring a password (KBA) or a token-based solution, Aculab’s voice biometrics engine (VoiSentry) uses a hybrid approach to analyse a voice sample as a means of highly accurate and secure authentication. This hybrid approach consists of:
- General phonemic, phonological and para-linguistic features as well as a number of unpublished proprietary analysis methods designed specifically for speaker recognition
- A combination of AI and further proprietary algorithms to combine the information from the various biometrics into a single result.
This highly-optimised combination of biometrics markers provides a much more accurate and reliable result than that of identification using just one or two markers.
These markers are compiled using AI algorithms to form a customer’s unique voiceprint. This process is very simple from an end-user’s perspective, with the user having to repeat a short phrase 3 times. Unlike other similar products, VoiSentry allows each user to create their own unique pass-phrase. This enhances security even further as any attacker would need to copy both the target voice and the spoken words to access. Once this voiceprint has been created it is held securely within the VoiSentry engine.
The next time the customer calls, it is as simple as comparing their voice on the call to their voiceprint to securely confirm their identity.
Not convinced? Try Aculab’s free demo to see if you can try fool the system. We wager you won’t.
How to implement these services?
As highlighted, traditional methods of ID&V are costly, inefficient, and place barriers between customers and their queries.
Voice Biometrics is truly the future. In a world where personal data is becoming less and less secure, there is one option that is virtually impossible to fake, hack or duplicate when combined with personal knowledge – our voice.