The next generation of Conversational Agents uses the latest Gemini models, enabling human-like, high definition voices, comprehension, and the ability to understand emotions — so the AI agents can adapt during conversations.

The product supports streaming video, so the AI agents can interpret and respond to what they see in real-time from customer devices. It will also simplify how AI agents are built, leveraging our latest Gemini models and Agent Development Kit, with a no-code console.

Users will have a variety of tools to interact and perform specific tasks with other applications (like add to cart, and checkout) through API calls.

Read this blog to learn more.

Our latest voice models provide human-sounding emotions and intonations, with steerable voice instructions and configurable pauses, pronunciation, and emotion to provide the life-like conversational experiences. 30 new voices with various tones to choose from are available now.

These quickstart reference agents allow users to get started easily with consumer commerce.

We are initially launching four prebuilt agents: flight booking, movie ticketing, shopping assistance, and appointment booking. They demonstrate various capabilities, including geolocation API use, connecting to Google products like BigQuery for data retrieval, Google Calendar for appointment booking, and rich-text formatting to build beautiful interfaces for end users.

We are making it easy to expand AI agent capabilities by introducing 30+ data retrieval connectors to augment the knowledge of your agent, and 70+ action connectors to allow your agent to take actions with or without human-in-the-loop for approvals.

Connectors include common data sources like BigQuery, Salesforce, SharePoint, Jira, ServiceNow, and more. With the new Conversational Agents connectors, customers get out-of-the-box, click-to-deploy tools that speed up time to production, provide seamless customer personalization, and quickly help realize the business impact of Conversational AI.