Google Launches New Gemini Models to Upgrade Enterprise Voice Agents

Google has introduced two new AI models, Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, to enhance its enterprise voice agents, the company announced on Tuesday (Sept. 15). The new models prioritize scale, cost efficiency, and high-complexity task handling, with pricing competitive to other frontier models. Both models offer advancements in near real-time reasoning, enabling more intuitive and intelligent conversations with AI.

According to Google, the new models provide developers and enterprises with building blocks for production-ready voice agents, enhancing the experience of interacting with Gemini across Gemini app, Google Workspace, and Search. PYMNTS previously highlighted the growing importance of voice technology in agentic commerce, with investments accelerating in voice-native AI infrastructure over the past year.

Voice is increasingly seen as middleware between large language models and end-user transactions, reducing friction for consumers ready to make purchases. PYMNTS reported in July that voice AI is evolving from answering questions to taking action, with companies like Anthropic and OpenAI deploying voice and chat agents for customer support and sales.

Google has also been expanding its voice AI offerings, launching new tools for Gmail, Google Docs, and Google Keep. These tools allow users to employ conversational search, engage in real-time conversations for document creation, and structure ideas through spoken commands, turning thoughts into organized lists and notes. Google emphasized that as technology advances, the way we complete daily tasks will continue to evolve, aiming to simplify tasks using just voice commands.