Yobe, a provider of intelligent voice technology, has launched VIENNA (voice intelligence extraction for neural network analytics), a voice extraction technology the company says makes it easy to extract clear, valuable insights from conversations, even in noisy environments. The new technology builds on Yobe’s software, which is based on AI and signal processing science developed by Yobe’s Chief Scientist and co-founder, Professor Hamid Nawab, which extracts meta-data, enabling machines to focus on person-specific audio.

The technology is designed to solve the long-standing challenge of isolating and transcribing speech in complex, real-world settings, using a blend of voice biometrics, intelligent voice activity detection and large language models (LLMs), to capture and analyze only the most relevant speech, filtering out background noise and irrelevant chatter.

The system even captures emotional cues like frustration or satisfaction, helping businesses gain deeper insights into customer sentiment, improve operational efficiency and deliver enhanced user experience. 

Companies can use VIENNA not only for tasks like identity verification and improving human-machine interaction but also to drive new product development by analyzing customer feedback and preferences in real time.

“Voice technology has come a long way— the latest breakthroughs with LLMs help machines to understand what we mean, not just what we say, but we still need to help machines listen to human voices, to distinguish voices from background noise and each other,” said Daren Gill, Member of Yobe’s Board of Directors and former product leader at Google, Spotify and Amazon.