Angelo Gajo | January 8, 2020 | 6 minutes
Integrating voice biometric capabilities into any business solution should be easy and hassle-free. However, security managers often spend countless hours tuning the settings and thresholds to obtain optimal security performance. Opting for a system-wide security setting can pose security issues as each speaker will have different levels of security, resulting in ineffective security performance. Auraya’s next-gen ArmorVox voice biometric engine and its patented features, allows organisations to easily deploy voice biometrics seamlessly and quickly. ArmorVox uses machine learning to optimise the security performance of every individual voiceprint automatically. Security managers can be assured of having the best security performance from day one of deployment with continuous optimisation for every user every time they use the system.
With ArmorVox, the security thresholds of each voiceprint are automatically calibrated using machine learning to create speaker-specific security thresholds. By calibrating each speaker’s security threshold levels, organisations can ensure a consistent level of security for every individual in the system. This means that organisations can truly achieve their desired False-Accept Rate (FAR) targets and satisfy any security requirements.
ArmorVox provides tools to allow security managers to simulate a massive hack attack where lots of enrolled users can be used to attack lots of other enrolled users. Using built-in tools to audit system performance, organisations can rest assured that their security is being constantly maintained at specified levels. The audit tools allow security managers to certify that the security thresholds that are nominated for different types of transactions are in fact being delivered by the solution.