Pindrop delivers improved audio intelligence to combat voice fraud

Although we tend to focus on frauds as a result of online account takeovers, more traditional social engineering methods are still a major problem, as are newer threats like deepfakes.

Voice technology company Pindrop is using this week's RSA Conference to launch new features that boost the level of intelligence that can be gained from voice analysis.

To improve authentication of users, Pindrop has new features that can predict the age range and spoken language of a caller, monitor fraudulent attempts to pass voice verification, as well as offering customizable account data tags for better policy automation.

"These do not have any dependency on previous interaction," says Amit Gupta, VP of product management, research and engineering at Pindrop. "In fact, one of the key advantages of this is that this information is available for every call that comes into the contact center for our customers, whether it's the first time a caller is calling in or whether it's the tenth day. Because what we are able to do in our technologies, we are able to extract specific acoustic features of the voice as the speech is detected on the line. And then we have trained machine learning-based engines to start looking at those acoustic features and based on that we developed these age prediction and gender prediction engines that run in the background and start making those predictions."

Beyond the additional intelligence extracted from voice interactions, the Pindrop solution also provides customers with custom data fields or tags that can be used to enhance policy creation -- providing more impactful call fraud investigations based on more intelligence provided by the contact centers directly.

Prior to this most recent release, Pindrop has partnered with Google to start offering its solutions to contact centers that leverage Google's Cloud Services.

You can find out more on the Pindrop site.

Image credit: Syda_Productions/depositphotos.com