Deep Voice™ Biometric Engine

DEEP VOICE™

FOR SPEAKER IDENTIFICATION

Pindrop’s proprietary, text-independent speaker recognition system passively identifies legitimate and fraudulent voices.

WHAT IS DEEP VOICE™ BIOMETRIC TECHNOLOGY?

The Deep Voice^™ engine is Pindrop’s proprietary deep neural network speaker recognition system. Akin to the deep neural networks that power autonomous cars, Deep Voice^™ biometrics runs throughout the lifecycle of a call, from IVR to agent, identifying and analyzing repeat callers. Running in the background of every call, the system offers friction-free voice analysis to authenticate legitimate callers and identify fraudsters.

AVOID BACKGROUND NOISE INTERFERENCE

Fraudsters attempting to mask their identity will often try to hide behind distracting or loud background noise. Pindrop Voiceprinting™ technology is designed to minimize sensitivity to background noise, reducing distractions and taking the advantage away from fraudsters.

DEEP ANALYSIS FROM SHORT UTTERANCES

Every caller's voice presents unique acoustic and behavioral features over the wire. Pindrop analyzes these unique signals, extracted from short utterances of a caller's speech — whether they are interacting with the IVR or with an agent — to develop a unique voiceprint for each caller.

VOICE SECURITY

Pindrop doesn't stop at the first enrolled call. As more calls are analyzed per caller, the credential grows stronger. This maturation process also protects against degradation of the voice over time.

HOW CALL CENTERS USE DEEP VOICE™ BIOMETRICS

Biometrics is the measurement and analysis of a person’s unique physical and behavioral characteristics. Call centers use Pindrop Voiceprinting™ technology to invisibly analyze characteristics of their customers’ voices. With the Deep VoiceTM engine, they can verify that callers are who they claim to be, or if a caller’s voice matches that of a known fraudster profile. Fraudsters beware: with Pindrop’s fraud detection solution, call centers can routinely update and indefinitely store “fraudster voiceprints” so trying to change things up or just laying low for a while won’t work!

WHITELIST

Compare a caller’s voice against an enrolled voiceprint tied to an identity claim.

BLACKLIST

Crossmatch a caller’s voice against Pindrop’s voice blacklist to detect repeat fraudsters.

ANOMALY DETECTION

Examine speech for voice morphing, recording and playback attacks, and simulated voices.

WHY OUR TECHNOLOGY IS DIFFERENT

Pindrop’s Deep Voice^™ biometric engine is the world’s first end-to-end deep neural network-based speaker recognition system. Whether callers are engaging with an IVR or speaking to an agent, it offers more accurate authentication in less time, using the same technology that powers autonomous vehicles and your favorite search engine. We built this proprietary engine from the ground up with an emphasis on security. Our voice biometric technology removes background noises from a call, so we focus solely on the caller’s voice — allowing us to identify attacks ranging from voice morphing and recorded voices, to synthesized and simulated voices.