Full Breakdown
Google Introduces Gemini 3.8 Live Avatar for Enterprise Customers
By Drooid · · How we work
Core Event
Google has added a “Live Avatar” to its Gemini 3.8 Live model, allowing enterprise users to converse with an animated AI persona that lip-syncs and displays facial expressions in real time. The feature is currently limited to Gemini Enterprise customers and is delivered through an allow-list process.
Background and Recent Model Release
The Live Avatar follows the launch of Gemini 3.8 Live and Live Extended Thinking models, which Google promoted as its most advanced live-dialogue systems. Those models were introduced a week earlier and are described as processing visual inputs in near real-time, providing the foundation for the new visual persona.
Features and Technical Capabilities
- Multilingual support: The avatar can operate in any of the 97 languages supported by Gemini 3.8 Live without visual degradation.
- Real-time lip-sync and expressions: Video demonstrations show the avatar’s mouth movements aligning with speech in both English and Japanese, and the system can display realistic or cartoon-style faces.
- Tool integration: While maintaining a visual presence, the avatar can trigger background tool calls and fetch data, enabling continuous dialogue.
- Customization: Google offers a library of preset avatars and allows organizations to create custom avatars from high-quality reference images to preserve brand styling or character identity.
- Transparency safeguards: Each output includes an invisible SynthID watermark, and Google states the system is built with safeguards designed to respect identity and keep AI-generated content transparent.
Official Position and Safeguards
Google emphasizes that the Live Avatar is intended to give developers and enterprises reliable, production-ready voice agents for tasks such as sales support and customer service. The company notes that the watermark and identity-respect safeguards are integral to the offering and that access is restricted to enterprise customers through an allow-list.
Potential Enterprise Impact
By pairing near-real-time visual presence with Gemini’s advanced reasoning, the Live Avatar aims to make multimodal conversations more natural, allowing avatars to convey information through facial cues while listening and speaking. If adopted widely, the technology could appear in customer-service interfaces on computers or phones, though its reception may vary among users.
