Illuma Answers Your Questions about Voice Biometrics

Voice biometrics is becoming a popular option for identity verification because it offers a convenient way to counter the threat of identity theft reaching new levels every year. Having new and better ways to protect personal data and accounts is more important than ever.

What Is Biometric Voice Recognition?

Biometric voice recognition is a method of verifying identity using the unique characteristics of a person’s voice. It relies on the fact that, like fingerprints, no two human voices are identical. Unlike speech recognition, which identifies the words a person is saying, voice recognition verifies whether or not the person speaking is who they claim to be. 

How Does Voice Biometrics Work?

Voice biometrics software works by extracting unique characteristics of a speakers’ voice and matching it to an enrolled sample stored securely in a biometric database. This previously collected sample may be referred to as a “voiceprint”, indicating its similarity to a fingerprint. An algorithm is used to compare a speaker’s voiceprint with their enrolled voiceprint to check for similarity. The level of match must be extremely high for the speaker’s identity to be verified.

How Accurate Is Voice Biometrics?

Voice biometrics accuracy is very high. The most important goal for voice verification is to keep an unauthorized user from accessing private information. This means it is most important to avoid false positives (incorrectly validating a fraudster). Tolerance for false negatives (incorrectly indicating that a caller is NOT who they claim to be), is a little higher since there are additional or secondary verification steps that can be used to validate an individual in these cases. 

A well-designed voice verification platform makes it possible to set the tolerances at a level that effectively weeds out fraudsters while limiting how many legitimate callers go through additional verification steps. Overall, the accuracy of voice recognition biometrics is extremely high, delivering very few false negative results when appropriately calibrated. Machine learning and AI can help the biometric voice authentication system to refine its results over time, becoming more and more accurate with each interaction.

How Secure Is Voice Biometrics?

Biometric voice recognition is highly secure, because the unique characteristics of an individual voice are unique and complex. Even closely related individuals have distinct voice patterns that are easy for a properly designed algorithm to differentiate. In addition to liveness testing, some versions of biometric voice recognition software can also incorporate unique characteristics of the individual’s calling device into the matching algorithm, providing an additional layer of security against recordings.

The privacy and security of voice recognition is also related to how and where biometrics are stored and how they are used in matching. For example, rather than storing a recording of a voice saying a passphrase, a more secure approach is to generate a proprietary and irreversible voiceprint or AudioPrint™ by using an algorithm to extract key characteristics of the voice and store it in a format that can’t be reverse engineered back into human speech. 

What Is Active Voice Biometrics?

Active voice biometrics requires the speaker to say a specific, preselected phrase for the voice matching to work. For example, on enrollment the user would be prompted to say, “My voice is my password.” This exact sentence would be their spoken password going forward. In contrast, passive voice recognition works by matching the speakers’ voice based on vocal patterns regardless of the words (or language) spoken. With passive voice verification, matching occurs in the background during a normal conversation. 

(Here’s a deeper dive on Active and Passive Voice Biometrics.)

Where Is Voice Biometrics Used?

Voice biometric technology can be deployed across any environment that utilizes a voice channel. This includes call center telephony platforms and digital voice channels including virtual agents, mobile, and IoT devices. For example, voice biometric solutions are frequently used for voice authentication in call centers where a live agent relies on the software to verify that a caller is who they claim to be before proceeding to address their question or request. Voice verification software can also be used in the IVR to help callers securely and conveniently access their account information and conduct transactions without speaking to an agent. 

Common Voice Biometrics Use Cases In Industry

Voice biometrics in banking is the most popular use case, since giving the wrong caller access to a legitimate consumer’s account would result in fraud loss, damage to the consumer relationship, and harm to the organization’s reputation. Because of this, the financial sector has been an early adopter of voice biometrics applications for fraud prevention.

Healthcare is another sector that can benefit from using voice verification to protect health records from unauthorized access. With the rise in telehealth consulting, emerging technology such as voice biometric technology will continue to grow in popularity.

E-commerce proves the utility of biometric voice authentication in self-serve digital channels. With intelligent virtual assistants or “Alexa” style IoT devices that verify the identity of the speaker, consumers can be confident that no one else can place a retail order using a voice activated device that is linked to their online store account.

What Are the Benefits of Voice Biometrics?

The advantages of voice recognition biometrics include greater security, operational efficiency, and a seamless consumer experience. Because a voiceprint can’t be effectively spoofed or hacked, it is very difficult for fraudsters to defeat a voice authentication system. Unlike security questions that can be overcome with answers stolen, guessed, or obtained through social engineering, voice verification is immune to these common threats. 

(Explore real-world results for a financial institution deploying voice authentication.)

How Does Voice Biometrics Work in a Call Center Environment?

With a voice biometric system such as Illuma Shield™, on answering a call the agent can immediately launch the voice verification software in the background and continue to speak with the caller. Within less than 15 seconds of natural conversation, the system will provide a positive or negative result on the caller’s identity. In this way, biometric voice authentication “recognizes” an individual’s voice and allows the agent to treat them warmly on every call. 

By removing time consuming and frustrating Q&A that comes with knowledge based authentication (KBA), consumers can have their needs serviced quickly. In a call center environment where consumers have already waited on hold to speak to a live agent, creating a frictionless verification pathway can substantially improve the caller’s view of the experience. Because a voice verification also takes less time than security questions, average call handle time decreases and efficiency rises when a voice authentication solution is deployed.

How Much Does Voice Biometrics Cost?

Voice biometrics companies can charge less than $25K to well over $1M. The wide range in pricing among voice biometrics providers often aligns with a difference in product category (e.g., active vs passive) and complexity, but not necessarily in quality. A right-sized approach to voice biometrics software focuses on providing state-of-the-art capabilities in a platform that is easy to deploy for small to mid-sized organizations.  

To learn more about proprietary voice biometrics from Illuma, contact us today.