Google Gemini 3.8 Live With Live Avatar Brings Real-Time AI Conversations

Google is taking its Gemini AI experience into a more visual direction with the launch of Gemini 3.8 Live with Live Avatar, a new enterprise feature that combines real-time voice conversations with a dynamically generated visual avatar. The feature adds near real-time video generation, lip-syncing and expressive facial movements to Google’s latest live dialogue technology.

Announced on September 24, 2026, Gemini 3.8 Live with Live Avatar is designed primarily for enterprises building conversational AI experiences for customers, employees and other users. Google says the technology can listen, see and speak while maintaining a continuous visual presence during a conversation.

The announcement follows Google’s launch of Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking earlier this month. Those models introduced faster and more capable real-time voice interactions, including visual understanding, multilingual conversations and background tool execution.

What is Gemini 3.8 Live with Live Avatar?

Gemini 3.8 Live with Live Avatar is Google’s enterprise AI technology that combines real-time voice conversations with generated video avatars, synchronized speech, facial expressions and support for 97 languages.

What Is Gemini 3.8 Live With Live Avatar?

Gemini 3.8 Live with Live Avatar is Google’s new conversational AI experience that adds a visual avatar to Gemini’s real-time voice capabilities.

Instead of interacting with an AI system through text or audio alone, users can communicate with an animated digital character that responds using synchronized speech, facial expressions and visual movements.

Google says the technology combines near real-time video generation with its native live dialogue models. The result is intended to make AI interactions feel more natural and engaging, particularly for enterprise applications such as customer service, guided experiences and interactive assistance.

The feature is currently available in Gemini Enterprise, with Google positioning it as a technology for organizations that want to build more interactive conversational agents.

Real-Time AI Conversations With Visual Presence

One of the main features of Gemini 3.8 Live with Live Avatar is its ability to maintain a continuous visual presence while processing a conversation.

The system can process audio and visual inputs simultaneously and generate corresponding audio and video responses. Google highlights natural turn-taking, expressive reactions and precise lip synchronization as important parts of the experience.

This could make conversational AI more suitable for situations where users expect an interactive digital representative rather than a traditional chatbot.

For example, an enterprise could use an avatar to guide a customer through a service process, explain how a product works or provide assistance during an interactive walkthrough. Google also demonstrates scenarios involving customer service and hotel check-in workflows.

Gemini 3.8 Live Can Handle Tasks in the Background

Another important capability comes from Gemini’s asynchronous tool execution.

According to Google, Live Avatar can make tool calls and retrieve information in the background while continuing the conversation. This means an AI agent does not necessarily have to stop talking while it waits for an external system to complete a task.

For enterprise applications, this could be useful when an AI agent needs to access databases, retrieve information or perform actions through connected services.

Google gives the example of a hotel check-in interaction where the AI can continue engaging with the user while background operations are being completed.

This builds on a broader capability introduced with Gemini 3.8 Live. Google’s September 15 announcement said the model can execute tools and API calls in the background while maintaining an ongoing conversation.

Supports 97 Languages

Multilingual communication is another major part of Gemini 3.8 Live with Live Avatar.

Google says the feature supports 97 languages and can transition between languages during a conversation. The system synchronizes speech, lip movements and facial expressions as the language changes.

This is particularly relevant for global businesses that need conversational AI capable of handling customers who switch languages during an interaction.

Google says the system is designed to maintain visual quality during these transitions without introducing visual drift or degrading the avatar’s lip synchronization.

The multilingual capability also builds on Gemini 3.8 Live’s ability to automatically detect and transition between 97 supported languages during live conversations.

Custom AI Avatars for Businesses

Google is also giving organizations the ability to create customized Live Avatars.

Alongside a selection of preset avatars, businesses can provide a high-quality reference image to generate an animated avatar. Google says this can preserve elements such as the reference likeness, brand styling or character identity.

Custom avatar creation is currently available through enterprise allowlisting, meaning it is not an unrestricted feature for every Gemini user.

This could allow companies to build AI representatives that better match their existing branding instead of relying exclusively on generic digital characters.

For example, a company could potentially use a branded avatar for customer support, employee onboarding or an interactive information kiosk.

Gemini 3.8 Live and Extended Thinking

The Live Avatar announcement is part of a larger Gemini 3.8 Live rollout.

On September 15, Google introduced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, describing them as its latest models for near real-time voice interactions. The standard Gemini 3.8 Live model focuses on scale, cost efficiency, fluid dialogue and visual grounding, while the Extended Thinking model is designed for more complex reasoning tasks.

Gemini 3.8 Live can process visual information during conversations and perform background tool calls. The Extended Thinking version goes further by allowing reasoning to continue while the AI speaks, helping it handle multi-step tasks without forcing the conversation to pause completely.

Google has also made the models available through the Gemini API and Google AI Studio for developers. Gemini 3.8 Live is also being rolled out across products including Gemini, Search and selected Google Workspace experiences.

Focus on Enterprise AI Agents

Google is positioning Gemini 3.8 Live with Live Avatar as an enterprise technology rather than simply another consumer chatbot feature.

The company says the technology can be used for customer service, interactive walkthroughs and other conversational applications where visual presence can improve the experience.

Google Cloud has also announced that Gemini 3.8 Live with Live Avatar is generally available in Gemini Enterprise. The company describes it as a platform for conversational video agents across web, mobile and interactive kiosk experiences.

This puts more emphasis on AI agents that can interact with people continuously rather than simply responding to individual text prompts.

AI Safety and SynthID Watermarking

Google says transparency is also a key part of the Live Avatar system.

AI-generated output from Google’s products is watermarked using SynthID, an imperceptible watermarking technology designed to help identify AI-generated content.

For Live Avatar, Google says SynthID is incorporated into generated audio and video. The company says this is intended to make AI-generated media detectable and help reduce the risk of misinformation and misattribution.

This is increasingly important as AI-generated video becomes more realistic and conversational avatars become easier to deploy.

When Is Gemini 3.8 Live With Live Avatar Available?

Google announced Gemini 3.8 Live with Live Avatar on September 24, 2026, and said it is available in Gemini Enterprise.

Developers and organizations interested in building with the technology can access Google’s API documentation and enterprise AI services. Custom avatar creation remains restricted through enterprise allowlisting.

For consumers, the broader Gemini 3.8 Live technology is appearing through Google’s Gemini ecosystem, including Gemini Live and Search experiences, but the Live Avatar feature is currently positioned specifically for enterprise use.

Why Gemini 3.8 Live With Live Avatar Matters

The launch shows Google’s broader direction for conversational AI. Instead of limiting AI interaction to text boxes, voice assistants or static interfaces, Google is building systems that can combine speech, vision, reasoning and generated video in a single conversation.

Gemini 3.8 Live already allows AI to understand visual context, switch between languages and perform background actions while maintaining a conversation. Live Avatar adds another layer by giving those interactions a visible, animated presence.

For businesses, this could create new applications across customer support, training, onboarding, education and interactive digital services. At the same time, Google’s use of SynthID highlights the importance of transparency as generated video becomes increasingly realistic.

With Gemini 3.8 Live with Live Avatar, Google is moving conversational AI closer to an experience where users can see and hear an AI agent respond in real time, rather than simply reading its answers on a screen.

Source: Google