Google has introduced Gemini 3.8 Live with Live Avatar, adding near real-time visual presence to its advanced live dialogue models. The feature pairs video avatars with voice agents to create an experience that listens, sees, and speaks with a dynamic visual persona. These avatars are designed for tasks like sales support and customer service, offering precise lip-syncing, natural expressions, and fluid turn-taking.

They are backed by Gemini's reasoning capabilities, allowing them to trigger tool calls and fetch data during conversations. Google will offer a library of diverse, preset avatars, while organizations can customize their own from high-quality reference images to preserve likeness and brand styling. The avatars are available only through enterprise allowlisting and include strict safeguards, such as Google's Synth ID watermarking, to ensure transparency.

While initially limited to Gemini Enterprise customers, these avatars may soon appear in customer service or sales roles on personal devices. Their reception may vary, given public skepticism toward AI-generated content.