Google Launches Gemini 3.8 Live and Extended Thinking for Voice-Driven Business Tools
Google introduced two new AI models under its Gemini 3.8 lineup, designed to improve voice interactions for businesses. The Live model offers scalable conversational intelligence, while the Extended Thinking version handles complex reasoning tasks. According to Google, the Extended Thinking model scored highly on speech quality benchmarks, potentially helping small firms cut customer service costs.
The two models serve distinct operational needs. The Live model prioritizes scalable, natural dialogue for customer-facing voice agents, while the Extended Thinking variant tackles multi-step reasoning, such as financial report generation. Google reports the Extended Thinking model achieved an 82.6 overall rating on its Speech to Speech Quality Index and a 97.7% score on Big Bench Audio, indicating strong performance in both conversational quality and complex reasoning.
Integration is designed to be seamless across Google's ecosystem, including Workspace and Search. The Live model is noted for its affordability, making it accessible for budget-conscious firms. However, adoption requires investment in technology and staff training, and businesses must address data privacy and regulatory compliance to avoid security pitfalls.
The introduction of these AI voice models could significantly reshape small business operations, potentially lowering customer service costs and freeing staff for strategic work. However, this shift may also create a digital divide, where firms without the resources for training or compliance struggle to keep pace. Customers could experience faster, more consistent service, but may also face frustration if AI fails to handle nuanced queries. Ultimately, the impact hinges on how responsibly businesses implement these tools, balancing efficiency with human oversight and data protection.