Research
We’ve been leading the way in spoken language technology and Conversational AI for more than four decades, and we’re not done yet.
Our research areas
We conduct research with our clients in mind. Learn more about each of the areas we focus on.
Acoustic Attributes
We conduct research into acoustic attributes to improve accuracy of authentication through voice biometrics technology, and decipher emotion and tone of the user for greater intent understanding.
Automatic Speech Recognition
Our customized Automatic Speech Recognition technology, including speech analytics, redactions, and real-time recognition, is built in-house and its models are constantly optimized.
Dialog/Conversational AI
At the core of our Dialog and Conversational AI research is an adaptive capability to recognize user intent through the combination of Conversational AI technologies with human intelligence at real time, allowing customers to speak naturally.
Human-Assisted AI
We were the first to use human-assisted AI, and the only to do it in real-time. We conduct research to further optimize human expertise to guide our technology without the knowledge of the user.
Machine Learning/Deep Neural Networks
Our processing includes proprietary Machine Learning capabilities, including deep neural network (DNN), support vector machine (SVM), MaxEnt, reinforcement learning, sequence learning, and more. The appropriate ML algorithm is used based on the data and task at hand to constantly improve AI models while reducing the number of utterances going to human assisted understanding.
Multi-modal Interactions
Our Multi-modal Interactions synthesizes information from multiple channels to fuse into a rich information source and interpret caller intent in a specific context. The result is a seamless, predictive omnichannel experience for customers.
Natural Language Processing
We have set the standard with the rich history of our Natural Language Processing technology to facilitate the most accurate interpretation of our client’s customer’s intent. The technology is language agnostic and can be adapted to multiple domains with minimal supervised data.
Voice Biometrics
Our Voice Biometric technology improves user experience by allowing active and passive authentication through I-vectors, X-vectors, and audio embedding features.
Our team
Behind every impressive technology stack is an even more impressive team. Our world-class research team creates technology that billions of people have interacted with. See what they’re up to.
Meet the Team