Lip-Sync and Asynchronous Execution: The Evolution of AI Agents with Corporate Faces
Google officially introduced Gemini 3.8 Live with Live Avatar, a new feature designed to move artificial intelligence interactions into an eminently visual environment. The system allows AI agents to respond through animated characters capable of lip-syncing, adapting their facial expressions, and maintaining real-time conversations.
From next+ we believe that the arrival of Live Avatar in Gemini 3.8 represents the consolidation of the "human interface" in corporate automation. Replacing traditional text chatbots with polyglot visual agents capable of querying systems in real time changes the dynamics of B2B and B2C customer service. However, the real value for companies will not only be in the visual appeal of the photorealistic avatar, but in the ability to securely integrate backend tools without generating friction. The mandatory inclusion of SynthID demonstrates that, in the era of hyperrealistic avatars, transparency about the synthetic origin of the interaction is the first condition for maintaining consumer trust.
Low-Latency Video Generation and Polyglot Fluency
Presented on September 24, 2026, the tool combines Gemini's real-time dialogue capabilities with low-latency video generation technology.
Among its technical capabilities are:
- Multilingual Synchronization: Supports 97 languages and can switch languages mid-conversation, adapting facial gestures without losing video fidelity or voice desynchronization.
- Turn-Taking and Gesticulation: Modifies facial expressions according to the interaction context to emulate a face-to-face conversation.
- Simultaneous Multimodal Processing: Analyzes audio and video streams at the same time to offer contextualized responses in a single interaction.
Asynchronous Tool Calls and Brand Customization
A key differentiator of the model is the ability to make asynchronous calls to background tools. This allows the avatar to consult databases or execute actions in external systems, such as verifying a guest's reservation during the hotel check-in process, while continuing to speak fluently with the user without pausing gestures or voice flow.
For business customization:
- Pre-configured and Custom Avatars: Includes a library of pre-designed characters and allows for the creation of custom avatars from reference images (subject to authorization list) to preserve brand visual identity.
- B2B Exclusivity: The feature is exclusively integrated into Gemini Enterprise for customer service channels, interactive tours, and corporate services, without being available to general users of the public Gemini app.
- Traceability with SynthID: Implements imperceptible digital watermark technology SynthID in generated audio and video, allowing identification of AI-produced content and mitigating risks of impersonation or misinformation.