Google Unveils Enhanced Voice Agent Models for Enterprise

Google has launched two new artificial intelligence models designed to power voice agents in enterprise settings. The Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking models are built for scale and cost efficiency, as well as high-complexity tasks.

The new models feature advancements in near real-time reasoning, making conversing with AI feel more intuitive and intelligent. They provide developers and enterprises with building blocks for production-ready voice agents.

These models enhance the experience of speaking with Gemini across various platforms, including the Gemini app, Google Workspace, and Search. The introduction of these new models follows a trend in investment in voice-native AI infrastructure, as reported by PYMNTS in February. Voice technology is becoming the foundational infrastructure for agentic commerce.