The success of remote work has pushed business communications to online meeting and chat tools such as Microsoft Teams and Zoom. This has brought about an opportunity for solving the long-standing language barrier in business communications since information now gets routed through servers capable of processing it. However, state-of-the-art natural language processing (NLP) models for audio transcription and translation are typically so large that it's impossible to make them process anything in real-time, during a meeting.

With HIVE™, we have developed a technology that enables building domain-specific compact versions of such high-accuracy NLP networks. HIVE™ can join Microsoft Tems meetings, listen to your conversations, transcribe them and translate them into any language, in real-time. Since HIVE™ can run completely on-prem, enterprises can utilize it to boost productivity in their internal communications as well as customer interactions.

HIVE™ also uses its transcription and translation features to support its conversational AI tools by ingesting meeting and call data for different use cases (e.g., Zoom calls with customers for sales and marketing use cases).