Nanosemantics Introduces Quorum Corporate Platform for Meeting Transcription and Analysis

Nanosemantics presented Quorum, a new corporate platform for automatic transcription, protocoling, and analysis of meetings. The solution is deployed on-premise in a closed corporate loop, so recordings and processing results do not leave the corporate infrastructure. Quorum processes audio and video recordings from Zoom, Microsoft Teams, corporate VCS, voice recorders, and other sources, turning them into a structured archive. The platform creates a transcript, a brief summary of the meeting, a protocol with tasks, and allows searching for information by meaning or asking questions to the archive in natural language. Quorum automatically identifies participants, highlights agenda, decisions, tasks, and responsible persons, and generates different types of summaries for executives, teams, or projects. Semantic search helps find relevant discussions even if the query does not match the wording in the recording. The platform runs entirely on the customer's servers, processing an hour of audio in about five minutes without transferring data to external services, with speech recognition accuracy exceeding 95%. The analytics and chat with the archive are based on a locally deployed LLM, ensuring summaries, tasks, and answers are generated without external data transfer. The platform adapts to company processes through corporate dictionaries, customizable protocol templates, and optional model fine-tuning, open API, and integrations with CRM, VCS, and internal systems. Nanosemantics estimates that using the platform reduces time spent on protocols and reports by up to 80%, and task capture accuracy reaches 95%. Dmitry Shulga, commercial director of Nanosemantics, highlighted that Quorum aims to transform chaotic verbal negotiations into a structured corporate asset, noting that 80% of important decisions are made in meetings but 50% of that information is lost. He emphasized that Quorum is the first system on the Russian market offering deep customization, RAG chat, and semantic search within a closed security perimeter.