Google launches Gemini Live Avatar for enterprise AI

Google has launched Gemini Live Avatar, a new feature for Gemini Enterprise that integrates real-time visual representations into conversational AI. The technology enables virtual agents to see, hear, and speak with high-fidelity video, featuring precise lip-syncing, natural facial expressions, and fluid turn-taking.

Designed primarily for enterprise use cases such as customer service, interactive tutorials, and digital advertising, the system can process visual and audio inputs simultaneously. This allows the AI to perform asynchronous tasks, such as managing a hotel check-in, without interrupting the verbal dialogue. The avatars support 97 languages, automatically adjusting facial movements and expressions when switching between them.

Organizations can choose from a library of preset avatars or create custom ones using high-quality reference images to maintain brand identity. To ensure transparency and prevent misuse, Google is applying SynthID digital watermarking to all generated audio and video content. The service is currently being rolled out to enterprise customers in the US and EU.