Voice biometrics: finding the most suitable method for your business

August 2, 2015 - 

This is a guest post by Graziela Barros, Product Manager for face and voice biometrics solutions at CPqD, in Brazil.

The use of voice biometrics to identify people is already a reality in the world: today, approximately 40 million people use this resource for commercial applications. Over the last two years, voice biometrics has outgrown the walls of academic innovation to become a real-world application, considered by companies all over the world as a powerful fraud prevention tool and more importantly, a way to improve relationships with their customers.

Biometric authentication – by voice, iris, fingerprints, hand veins or other physical characteristics – is much more secure than other traditional forms of authentication, such as passwords and security questions, since it is based on unique and exclusive individual attributes that cannot be shared. That is why it was first adopted in the defense and security segment. However, as this technology evolved, it began to gain important space in the market, such as the financial, telecommunications, government and health segments.

In the case of voice biometrics, the greatest benefit to companies is to provide their customers a more secure – and extremely natural – form of authentication. Every situation where a person is stimulated to speak represents an excellent opportunity for voice authentication.

Banks and carriers have already begun to use voice authentication in contact centers (phone calls) in several parts of the world, and customers have demonstrated increased satisfaction and assurance with this new process. Mobile applications have also begun to use the technology, allowing users to interact with the application entirely by voice, for a much easier and intuitive experience. This new form of interaction helps include the visually impaired, people with motor difficulties and those who are less technologically savvy.

Voice biometrics can be used in several different customer relationship channels, for an increasingly better user-experience. It is this flexibility that contributes to speed up the dissemination and the success of voice biometrics all over the world, since authentication methods can be adjusted to the desired usability of each situation, without compromising the expected security.

The use of voice biometrics can be implemented by different methods, always seeking the best balance between security and the most ideal user experience for each scenario.

One of the most well-known voice authentication methods is the “spoken passphrase,” where people memorize a sentence as their vocal password. When vocalizing this phrase – and only this one -, it is recognized by the characteristics of their voice and the content of the sentence. This method returns excellent accuracy rates with short audios, and is therefore one of the best options for interactive voice response (IVR).

By using this method, higher risk services can migrate to automatic customer service, since authentication is more reliable. However, there is a user resistance factor: they will need to remember a specific phrase every time they wish to do a transaction. In frequent access situations, the impact is minimal; however, in other situations, it is important to compare the gain against the cost of experience.

Graziela-Barros-voice-biometrics-chart-new

With this in mind, the method known as “positive identification/confirmation” deserves a closer look. When this method is used, people are authenticated by their voice during the traditional positive identification process already adopted by contact center operations. While vocalizing their personal data, people are submitted to biometric and knowledge evaluations (based on speech recognition technology). Users thus do not have to deal with unfamiliar paradigms, since they are already used to be positively identified by companies. Besides increasing the level of security of this process, with voice biometrics, the identification process becomes shorter and less bothersome, with fewer questions needed to validate a customer’s identity. And, the faster the service session, the lower the cost for the company and the greater their customer’s satisfaction. This method can be used for both IVR service sessions and human service sessions.

There are situations, however, when customers demand a fast response and it is not appropriate to bother them with a lot of questions. In this context, the “free speech” method is the most appropriate. In cases where users talk to a human attendant, without following a stimulated script, they can be authenticated in a transparent manner, with the entire audio of their call examined for biometric verification. This provides the most natural user experience, since the customer is merely talking and does not need to answer several personal questions.

Companies can adopt a different method for each one of their business units, for each transaction type, or for each client segment. The chosen solution must be able to provide alternatives to achieve the ideal balance between user experience and security for each operation, always considering the best result to the company.

DISCLAIMER: BiometricUpdate.com blogs are submitted content. The views expressed in this blog are that of the author, and don’t necessarily reflect the views of BiometricUpdate.com.

Leave a Comment

comments

About Graziela Barros

Graziela Barros is Product Manager for face and voice biometric solutions at CPqD, in Brazil. With a background in strategy consultancy, she believes that Technology is an important pillar of business strategy and that biometrics can greatly improve the relationship between companies and customers.