Auraya updates ArmorVox voice biometric suite with confidence scores
Auraya has announced the release of newly updated voice biometrics software suite ArmorVox to add a confidence scoring system and a serverless architecture option.
Providing cutting edge features like machine learning algorithms for active learning and imposter mapping, the software is also able to detect and repel synthetic voice systems.
ArmorVox Version 17.9.2 allows organizations to use an additional scoring system, logarithm odds scores, to reveal the relative probability of the accuracy of the voice biometric verification attempt. This system is supported by an updated calibration routine that models the true speaker scores in addition to the imposter score, providing human decision-makers and other AI systems with higher confidence when making identity verifications and identifying fraudulent activity.
AWS CloudFormation templates for AWS Cloud deployments are now supported, for simpler and lower-friction voice biometrics rollouts. AWS CloudFormation templates included with ArmorVox enable customers to quickly build load-balanced clusters.
The new version of ArmorVox will help to deploy voice biometric solutions faster with the option of serverless deployments based on AWS Lambda, helping to reduce costs and maintenance of traditional servers as well as providing efficient auto-scaling performance, according to the announcement.
The company announced earlier this month that EVA Voice Biometrics had been granted accreditation by the AWS Foundational Technical Review, meaning customers can use the technology on the Amazon Connect contact center platform.
Improvements to ArmorVox’s security include standard HTTP Basic Authentication on APIs and recording the voiceprint creation date will help the company’s network of partners to deliver secure solutions to customers.
ArmorVox has already been implemented by New Zealand’s Inland Revenue Department, Vodafone and ShadowTrack among other customers.