Zoom has unveiled AI Companion 3.0, a sweeping update that redefines the platform as a comprehensive, AI-first productivity ecosystem extending far beyond its video conferencing origins.
The release introduces a dedicated web surface for the AI assistant and, critically, democratizes access to core AI features for users on the free Basic tier. The company positions the AI Companion 3.0 as a solution for “conversation to completion,” aiming to eliminate the friction between discussion and actionable outcomes.
The AI Companion is now accessible via a new, permanent conversational work surface at ai.zoom.us on a desktop web browser, establishing the assistant as a central hub for daily work outside of live meetings. This expansion is supported by a strategic freemium model designed to drive adoption.
Register for Tekedia Mini-MBA edition 19 (Feb 9 – May 2, 2026): big discounts for early bird.
Tekedia AI in Business Masterclass opens registrations.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab: From Technical Design to Deployment (next edition begins Jan 24 2026).
Basic plan holders gain access to the AI Companion in up to three meetings per month for free. During these sessions, they can utilize high-value features such as meeting summary, in-meeting question answering, and AI note-taking. Additionally, free users can ask up to 20 questions each month via the side panel or the new web surface to retrieve highlights or action items from past meetings.
Full, unrestricted access to the AI Companion is available as a $10 per user per month add-on plan, which can be purchased without needing a separate paid Zoom Workplace license. An advanced Custom Companion tier is also offered for enterprise users, providing deeper customization, personalized knowledge collections, and integrations with their proprietary data sources.
Agentic Capabilities and Cross-Platform Orchestration
The most significant advancement in AI Companion 3.0 is its shift toward “agentic” capabilities, allowing it to perform multi-step actions and retrieve information across silos, turning scattered work conversations into continuous intelligence.
This intelligent assistance covers the entire workflow. The AI Companion features agentic retrieval capabilities, enabling it to pull information not only from all data stored within the Zoom ecosystem (meetings, chats, notes) but also from connected third-party platforms.
It currently supports Google Drive and Microsoft OneDrive, with planned, imminent support for Gmail and Microsoft Outlook, allowing the assistant to pull in email and document context for more informed responses. The system can even take notes for meetings held on Microsoft Teams and Google Meet, addressing the reality of mixed-platform work environments.
The assistant proactively manages the workday by generating a Daily Reflection Report, which summarizes meetings, tasks, and updates. It also automates post-meeting work with the Post-Meeting Follow-Up prompt template, which generates next steps, tasks, and drafts follow-up email messages. Custom AI agents, currently in beta for power users, allow a low-code design of personal workflows that automate routine tasks like summarizing chat threads every morning.
A new Agentic Writing Mode empowers users to draft, edit, and refine business documents using context derived directly from meeting discussions and documents. Users can start collaborative projects within the companion interface and seamlessly shift them to Zoom Docs, supporting exports to MD, PDF, Microsoft Word, and Zoom Docs formats.
Zoom, founded by CEO Eric Yuan, is actively competing with productivity behemoths like Microsoft (Copilot) and Google (Gemini) by leveraging its massive base of meeting data and an independent, platform-agnostic approach.
Lijuan Qin, head of AI product at Zoom, emphasized that the company’s independence and access to deep contextual meeting data give it a crucial advantage. Zoom’s technical core is its federated AI approach, which strategically combines the power of Zoom’s own custom Large Language Models (LLMs) and Small Language Models (SLMs) with the best models from third-party partners like OpenAI, Anthropic, and open-source models like NVIDIA Nemotron.
This hybrid approach dynamically routes tasks to the most suitable model for a given function, optimizing for performance, cost efficiency, and accuracy. This system has reportedly shown superior performance on certain benchmarks compared to reliance on a single frontier model.
By offering its AI assistant with limited free access and focusing on cross-platform functionality, Zoom is banking on a freemium strategy that captures new users and positions the AI Companion as an indispensable, neutral productivity layer across the entire enterprise stack, regardless of a company’s core software provider.