"As AI Companion's capabilities expand across new surfaces, Zoom stressed that it encrypts customer data in transit and does not use communications-like customer content to train its own or third-party models. AI Companion 3.0 uses Zoom's federated AI approach that combines its own large and small language models (LLMs/ SLMs) with third-party LLMs from OpenAI and Anthropic. Zoom users can also use open source models such as Nvidia Nemotron, which enables reasoning and retrieval-augmented generation within Zoom's federated AI stack, creating a personalised work environment."
"AI Companion has the stated aim of helping people start their day with clarity and alignment, even when dealing with numerous post-meeting tasks and follow-ups. Features of the system include a beta version of AI-first capabilities for personal workflows, agentic AI features for Zoom Docs, and a web interface with expanded context to help users uncover insights, optimise their day and improve their work."
AI Companion 3.0 introduces a conversational work surface that connects to work interactions to turn scattered meeting conversations and assets into insights, progress tracking, polished documents and content without requiring users to upload transcripts or craft detailed prompts. The release includes beta AI-first personal workflow features, agentic capabilities for Zoom Docs, and a web interface with expanded context to help users uncover insights and optimise their day. The system uses a federated AI stack combining Zoom's LLMs/SLMs with third-party models from OpenAI and Anthropic and open source models like Nvidia Nemotron. Customer data is encrypted in transit and communications-like content is not used to train models, while improved transcription, captions and translations aim to boost information retrieval and task completion.
Read at ComputerWeekly.com
Unable to calculate read time
Collection
[
|
...
]