Voice Biometrics has traditionally been implemented as a ‘point solution’ in contact centers and mobile devices. In our recent paper, “Voice Biometrics, What Could Go Wrong?“, we stressed the importance of multi-layer, multi-modal and multi-factor strategies.

Eckoh, the London-based provider of global payment products and customer service solutions, has been assisting organizations to de-scope their Payment Card Industry Data Security Standards (PCI DSS) compliance by masking sensitive card information that is entered via DTMF tones through their patented ‘audio tokenization’ CallGuard solution. They recently announced that they will be awarded a patent that combines voice biometrics with phone ‘footprinting’, adding yet another innovative layer to their secure payments solutions.

Other firms that have been offering various audio-path authentication include TrustID with their patented network, telephony and phone verification solutions, and Pindrop’s whose patented Phoneprint™ technology has been in place since 2015, and has already demonstrated immense value to curbing millions of dollars in contact center fraud for IVR and live-agent calls.

When one considers that a phone ‘footprint’ is also regarded as a unique template, very much like a biometric template (voice, finger, face, etc), this combination of factors actually satisfies all three criteria:

• Multi-layer – speaker and device, both components that make up the ‘edge’ of a connection
• Multi-modal – voice and phone ‘biometrics’
• Multi-factor – something you are (voice) and something you ‘have’ (phone)

The application of Big Data Analytics on various network components such as the SIM card (the primary network authentication layer), phone, carrier network(s), internal enterprise systems and other nodes across the entire voice ecosystem is yielding immense improvements in caller authentication. AT&T, Sprint, T-Mobile and Verizon announced in September that they have joined their considerable resources to create the Mobile Authentication Taskforce, whose goal to develop mobile authentication solutions using components such as network-based device authentication, geo-location and SIM card recognition.

As Machine Learning is applied to an ever increasing set of nodes across the entire voice ecosystem, we expect to see exponential improvements in Intelligent Authentication.

Categories: Conversational Intelligence, Intelligent Authentication, Articles