Phoenix 24
The interface is becoming part of the intelligence.
Mountain View
Google has introduced Gemini 3.8 Live with Live Avatar, a new enterprise feature that adds a synchronized animated face to real time artificial intelligence conversations. The system combines Gemini’s live dialogue capabilities with low latency video generation, allowing an avatar to listen, respond and maintain facial expressions and lip synchronization during an interaction. For businesses, the shift turns conversational AI from a disembodied assistant into a visible digital representative.
The tool is designed specifically for companies using Google Cloud rather than for general consumers. Organizations can deploy the avatar in customer service environments, guided experiences and other situations where a more humanlike interface could reduce the distance between automated systems and users. Google says the technology is already available through Gemini Enterprise, expanding the role of AI agents from text and voice toward visually embodied interaction.
One of the most important capabilities is what happens behind the conversation. The avatar can execute tools and retrieve information in the background while continuing to speak with the user, allowing the interaction to remain uninterrupted. Google demonstrated the concept with a hotel check in scenario in which the system verifies a reservation and completes internal procedures while maintaining the dialogue. That combination of visible presence and invisible task execution could make AI systems feel substantially more integrated into service workflows.
The global scale of the system is also significant. Live Avatar can synchronize speech, facial movement and expressions across 97 languages and can switch languages during the same conversation without interrupting the visual flow. Businesses can select from predefined avatars or, in authorized cases, create customized versions based on a reference image. This opens the possibility of building virtual representatives aligned with a company’s visual identity.
Google has also incorporated SynthID into the system, embedding an imperceptible watermark into the generated audio and video. The objective is to preserve traceability and make it possible to identify content created by artificial intelligence. As synthetic faces become more realistic, that distinction will increasingly matter for transparency, trust and accountability.
The deeper transformation is not simply that AI can now speak with a face. It is that companies are beginning to design artificial agents as visible participants in customer relationships. The next competitive frontier may therefore depend not only on what an AI knows, but on how convincingly it can occupy the space where humans expect another person to be.
Phoenix24: clarity in the grey zone. / Phoenix24: claridad en la zona gris.