Pindrop has unveiled the latest version of its Deep Voice voice recognition platform. The updated solution made its debut at this week’s RSA Conference in San Francisco.
Deep Voice was first announced in early 2018, and is designed primarily for contact center applications. Deep Voice 3 improves on the speed and accuracy of the original, enabling a 30 percent increase in authentication and enrollment rates. The new platform is also able to identify a caller based on short speech segments in as little as three syllables.
Pindrop believes that that last fact will give Deep Voice 3 even greater utility in IoT applications, which the company has been targeting since CES 2019. To that end, Deep Voice 3 can be used to identify people who are issuing simple voice commands to a virtual assistant like Alexa.
In the meantime, call center operators will benefit from the platform’s enhanced ability to deal with background noise during customer interactions.
“Confidently authenticating more callers, and doing so at a significantly faster rate, is taking us a step closer to frictionless identity in a voice-first world,” said Pindrop CEO Vijay Balasubramaniyan.
Deep Voice 3 is part of Pindrop’s Passport authentication platform, and will be available sometime in the current quarter. It will then be integrated into the company’s Protect anti-fraud offering within the next few months.
Pindrop’s release of Deep Voice 3 coincides with its release of an updated Voice Intelligence and Security Report, which found that one out of every 770 call center interactions was fraudulent in 2019. That rate is slightly lower than the report’s 1 in 685 rate for 2018. However, banks still experienced a 15 percent increase in phone fraud, which suggests that fraudsters are specifically going after high value targets.
Pindrop noted that that trend is particularly concerning because banking call centers have the highest volume of incoming calls.
February 27, 2020 – by Eric Weiss