Gemini Live Avatar makes enterprise AI agents look back
Google says Gemini 3.8 Live with Live Avatar adds near real-time visual presence to conversational AI, with speech, expressions, lip-sync, and tool use running together.
What happened
Google introduced Gemini 3.8 Live with Live Avatar, a Gemini Enterprise feature that combines live dialogue with low-latency generated video. The company says the system can listen, see, speak, and present a dynamic visual persona during a conversation.
The pitch is aimed at enterprise agents: customer service, walkthroughs, and branded virtual assistants that need more presence than a text box or disembodied voice.
The geeky part
Google says Live Avatar supports precise lip-syncing, natural expressions, fluid turn-taking, and native multilingual speech-to-speech synchronization across 97 languages. It also supports asynchronous tool execution, so an agent can call tools in the background while keeping the conversation going.
Custom avatar creation is available only through enterprise allowlisting, while preset avatars are offered more broadly inside Gemini Enterprise.
Quick takeaways
- Live Avatar is available in Gemini Enterprise.
- Google says SynthID watermarking is built into generated audio and video output.
- The feature is enterprise-focused, not a general consumer avatar product.