Spin-off Fluent.ai develops voice interface software which learns to understand the meaning of audible speech

Fluent.ai’s speech interface solution works in most languages and accents, allowing users to use their natural speech. Fluent.ai’s patented approach sidesteps the need to convert speech to text and instead extracts intent directly from speech. The technology was developed at Canadian and Flemish universities, among which are the University of Leuven (KU Leuven) and the University of Antwerp (UAntwerp).

Fluent.ai’s speech recognition technology addresses many of the problems of current voice interfaces, including their need for large amount of training data set and compute power. Fluent.ai’s solution results in a highly flexible, accessible and accurate voice interface that performs robustly even in offline and noisy environments. The technology’s ability to work accurately in any language and to perform well in noisy conditions represents a tremendous market opportunity, especially in the era of IoT and edge computing.

The patented acoustics-only solution learns the meaning of specific speech patterns and triggers actions accordingly. Because it doesn’t rely on standardized text-based vocabularies, it aligns to individual users, even those with complex languages, accents or voice impairments.

Fluent.ai’s software is based on artificial intelligence technology that was developed by Canadian and Flemish universities. Both the PSI (Processing of Speech and Images) Research Group of the KU Leuven Department of Electrical Engineering and the CLiPS Research Group (Computational Linguistics & Psycholinguistics) of UAntwerp delivered parts of the technology and contributed to this international spin-off company. Fluent.ai just raised CA$2 million in seed financing to accelerate commercialization of its AI voice interface technology.