Things to Consider When Selecting a Voice Biometric Vendor
All forms of Biometrics are being considered by organizations to improve security and privacy for their staff and customers. Voice biometrics can provide secure and frictionless identity verification and fraud detection capabilities for all customer channels. Voice Biometrics also provides benefits such as improving user experience, reducing costs, and making it easier for organizations to comply with regulations protecting privacy and money laundering (HIPAA, GDPR, KYC).
When considering a voice biometric solution, it is important to ask the right questions to gain more understanding about the benefits of the technology capabilities. Of the many questions, you would typically ask a voice biometric vendor, these are a few important questions that should be addressed.
1. How long does it take to deploy a voice biometric integration?
With EVA, the implementation can take anywhere between two hours to a few days, depending on the options selected for integration with agent desktops and existing CRM systems.
Organizations that have an existing AWS service can take advantage of tools such as Cloud formation to easily set up a highly available, highly secure, auto-scaling voice biometric solution where all customer data is retained and controlled by the client organization.
EVA from AWS Marketplace is a turnkey voice biometric solution for Amazon Connect. Customers. This service can be activated and start enrolling and verifying customers with consumption-based fees simply added to the AWS account at the end of each month. This consumption-based offer allows organizations to turn on a voice biometric capability and try it out without risk as the usage charges cease if the system is turned off.
2. What types of voiceprint tokens can be captured?
Auraya’s voice biometric technology can capture and process any number of text-independent, text-dependent, digit-dependent, or digit-independent voiceprints for an individual. For example, organizations can ask users to speak a digit-dependent unique phrase such as their customer account number or phone number. The digit-independent capability allows organizations to verify user identity by asking users to say a ticket number or an invoice number. This process provides a user-friendly way to identify the claimed identity and verify the identity biometrically with one spoken utterance.
Organizations can capture all types of voiceprints passively and actively in any channel. A person may enroll their voiceprint during an onboarding process on a website providing proof of identity documentation. The web enrolment of their voice can be used if they call the contact center or engage in a chat or messaging interaction.
The full range of voiceprint types increases an organization’s ability to provide personalized secure self-service in the customer's preferred channel.
3. Can users enroll and verify in any language?
With Auraya’s voice biometric technology, organizations can use Auraya’s voice biometric technology to enroll and verify users in any language. With advanced AI and machine learning algorithms, Auraya matches the voiceprint captured during interaction with the stored voiceprint in the database. As long as the captured voiceprint is said in the same language as the stored enrolled voiceprint in the database, organizations will be able to verify user identity in any language.
4. Can the voiceprints be updated over time?
With Auraya’s active learning capability, every time a user successfully verifies their identity, their voiceprint is tuned and improved. This means that every time a user is successfully verified, their voiceprint accuracy performance increases, making it even easier for the user to confirm their identity every time.
5. How do you protect against synthetic voice and recorded playback attacks?
Auraya uses machine learning algorithms to detect synthetic voice systems and recording attacks.
Auraya can also protect against recorded playback attacks. Auraya protects against recorded playback attacks by issuing random phrase challenges. If an organization suspects that there may be a recorded playback attack, an automated process for asking for a random phrase challenge is an effective process to thwart the attack.
6. Can voiceprints be used in different channels?
Auraya’s voice biometric technology can be implemented into any telephony or digital channel. It can be deployed into contact centers, web browsers, chatbots, mobile applications, and other platforms.
Additionally, the voiceprints can be used interchangeably between the different platforms. For example, organizations can enroll users via a web browser and verify them in the contact center via the IVR. This helps improve user experience and provides a smoother user journey for those who use multiple platforms to connect with organizations.
7. Where is the data stored?
Auraya technology allows an organization to choose where data is stored. They can store the data with a managed service provider or store it in their secure database on-premise or in the cloud. Wherever the data is stored, it is encrypted and cannot be reversed engineered to obtain playable audio. No matter what, the data doesn’t leave the organization’s control, and no one, not even Auraya, will be able to access the data without being granted access to the secure service.
There are many other questions organizations should ask voice biometric vendors, although these questions provide a good starting point. In addition to asking questions and obtaining more information, organizations can evaluate Auraya’s voice biometric technology via the Auraya Evaluation Portal. Today organizations can turn on a proof of concept or a full-blown production system by activating EVA from AWS marketplace.
If you’re interested in getting more information or want to evaluate Auraya’s voice biometric technology, contact us at info@aurayasystems.com and get started today!