← Back

Google's Gemini Voice AI Direct-Calls Businesses, Escalating Assistant Rivalry

Sep 25, 2026
Google's Gemini Voice AI Direct-Calls Businesses, Escalating Assistant Rivalry

Google is reactivating its controversial Duplex-style AI, enabling Gemini to make direct, voice-based calls to businesses for tasks like appointment booking. This move, years after the initial 2018 Duplex demo, strategically reclaims the narrative around agentic AI assistants, directly challenging Apple's Siri, which currently lacks comparable real-world task execution capabilities. By deploying this on-device, Google isn't just adding a feature; it's establishing a new baseline for what a mobile AI assistant must do, shifting the competitive focus from simple Q&A to complex, multi-step actions in the physical world. The system fundamentally alters the user-to-business interface, creating an asymmetric advantage for Google in local commerce data acquisition. Every call Gemini makes—confirming store hours, checking stock, or booking a table—is a structured data-gathering event, enriching its local search and mapping services with real-time, high-fidelity information that rivals like Yelp and even Apple Maps cannot easily replicate. While users gain convenience, small businesses become unwitting data sources. This forces a strategic recalculation for any platform reliant on local business listings, as Google begins to programmatically verify and update the world's commercial data at scale. The real test will be merchant and consumer tolerance for AI-intermediated communication, determining the feature's scaling velocity. Within six months, watch for competitors like Apple to announce deeper OS-level integrations for Siri at WWDC 2024 to counter this agentic leap. The critical variable is whether the convenience of delegating calls outweighs the inevitable errors and impersonality of AI interaction. This trajectory suggests a future where AI agents, not users, become the primary interactors with local businesses, reshaping the entire local search and discovery ecosystem within three years.