Imagine asking an AI assistant to find a customer, negotiate a price and help close a sale, all without typing a single prompt. Voice AI is moving beyond answering questions to taking on tasks, and Google Cloud wants businesses to put it to work.

At TechSparks 2026, Naren Kachroo, Head of AI Go-to-Market at Google Cloud, demonstrated how Gemini’s voice AI capabilities could help businesses handle customer conversations, restaurant bookings and sales negotiations.

Until now, much of the AI experience has revolved around asking questions and getting answers. Kachroo said the next step is AI that can access information, use tools and take action on a user’s behalf.

“We don't want to focus on voice AI just being an experience to talk to AI,” he said. “We want voice AI to be powered through agentic experiences where the task is actually accomplished with AI.”

To show what that means, Naren Kachroo played a demonstration of an AI assistant booking a restaurant table for eight. The assistant answered questions about bringing a birthday cake, identified a dish from an image, discussed vegetarian options and switched between English and Spanish during the conversation.

In other words, it did more than take a booking. It handled the questions that came with it, too.

Kachroo also introduced Gemini Live voice capabilities, alongside models for text-to-speech, speech-to-text and translation. He showcased Live Avatar, which combines voice interaction with a visual AI presence.

The standout demonstration involved a fictional safety equipment business with 500 helmets sitting in its inventory.

Source link