Google adds Live Avatar to Gemini 3.8 Live, live in Gemini Enterprise

Google's announcement pairs two things it already offers separately: a live conversational voice model and low-latency streaming video. By putting an avatar on top of that combination, Google is presenting the interaction as something closer to a face-to-face conversation, where the user can listen, watch and speak rather than just hear a voice.

The company is positioning the feature chiefly around customer-facing settings. It cites customer support and interactive walkthroughs as the kinds of digital exchanges the avatar is meant to make feel more real, and it lists accurate lip sync, natural expressions and smooth back-and-forth as the qualities that carry that effect.

The multi-language side is where the claim gets specific: Google says lip sync and expressions adjust dynamically so that switching among 97 languages happens without visual misalignment. The stakes for enterprise buyers therefore appear to hinge less on whether an avatar can be displayed and more on whether that alignment and expressiveness hold up in real service conversations, where a customer's trust in what they are seeing is on the line.