Google has moved Gemini 3.8 Live with Live Avatar from preview to general availability in Gemini Enterprise, following last week’s announcement. The feature was first shown at Google Cloud Next 2026 and now targets production enterprise voice-AI deployments. It combines the native speech-to-speech foundation of Gemini 3.8 Live with an interactive visual presence for conversational video agents on web, mobile, and interactive kiosks.

Available capabilities include video avatars with synchronized lip-syncing (custom avatars remain behind an allowlist), fluid dialogue with natural interruption recovery that preserves conversation context and backend transactions, asynchronous tool calling that runs API/CRM/ERP calls in the background while the agent continues talking, automatic language detection and switching across 97 supported languages mid-conversation, and live visual understanding that processes camera feeds and screen shares jointly with audio. Gemini 3.8 Live Extended Thinking is still in private preview. The GA offering is served from US and EU endpoints with provisioned throughput, enterprise compliance, and strict data governance.

To reduce misuse, deployment is limited to a curated library of pre-built avatars, and custom avatar creation requires a strict enterprise allowlisting and verification process. All generated audio and video streams include imperceptible SynthID watermarks for transparent content provenance.

Three demos illustrate the release: step-by-step custom avatar creation from system instructions plus a single reference photo and audio sample; an insurance claims intake agent that fills a claim notebook from live camera conversations while an ADK agent team checks policy and assembles an adjuster packet; and a developer workflow that uses Agent Development Kit (ADK) to define agents, manage runners and session memory, and stream real-time audio to the Gemini Live API without a traditional speech-to-text pipeline.

Customer mentions: Cox Automotive built an Autotrader shopping assistant with live screen-highlighting and tool calling for search, comparison, and financing; Equal AI says its personal AI handles over a million live calls daily across nine Indian languages and credits Gemini 3.8 Live with improved interruption handling, multilingual conversations, and tool-call reliability; Salesforce is combining Gemini 3.8 Live with Agentforce for real-time multimodal service; and Specs reported strong results in Voice Activity Detection and latency during benchmarking.