Google introduced Gemini 3.8 Live with Live Avatar, adding a real-time visual presence to its conversational AI. The new feature combines low-latency video generation with speech. Thus, allowing AI agents to see visual inputs, listen to users and respond through expressive audio and video. Google announced the feature on September 24, 2026.
Live Avatar builds on Gemini 3.8 Live, which Google launched the previous week. The system uses native speech-to-speech technology for more fluid conversations. With the new avatar layer, responses can include synchronised lip movements, facial expressions and different visual personas.
Google further explained, “ The technology is designed to make enterprise interactions more natural. Businesses can use the avatars for customer service, interactive demonstrations and other conversational experiences.”
Gemini 3.8 Live with Live Avatar can process audio alongside live camera feeds and screen shares. This allows an AI agent to understand what a user is showing while continuing a spoken conversation.
The tech giant demonstrated its capability through an insurance claims workflow. A customer can describe damage while showing it on camera. The AI can process the visual information while an agent system handles the claim workflow in the background.
Google says, “Gemini 3.8 Live can also automatically detect and transition between 97 supported languages. This allows multilingual conversations without requiring users to manually switch languages.” The system also supports asynchronous tool calling. This allows agents to make API, CRM, and ERP calls while continuing the conversation, reducing pauses during interactions.
Also Read: Google Gemini Expands Beyond Chat with 14 New Connected Apps