Google is not waiting for users to open the Gemini app. Instead, it’s putting the AI directly inside Chrome, where people actually spend most of their time on their phones. The company announced that Gemini in Chrome is now available to all Android users in the U.S., and the most interesting part isn’t the summarization feature. It’s auto browse, an agentic mode that can actually take actions on your behalf.
Auto browse is limited to AI Pro and AI Ultra subscribers, which puts it in the same tier as Google’s premium Gemini Advanced offerings. The feature can handle tasks like booking parking, updating a recurring online order, or organizing travel itineraries. That puts it in direct competition with tools like Perplexity’s new agentic search, OpenAI’s Operator, and Anthropic’s computer use. But the difference here is context. Gemini in Chrome operates inside the browser you’re already using, rather than asking you to open a separate interface and hand over control.
The broader feature set available to all users includes page summarization, question answering about web content, and integration with Google Calendar and Keep. That last piece matters more than it sounds. Cross-app functionality without tab switching is exactly where mobile AI has struggled, and Google has a structural advantage here because it owns both the browser and the productivity apps.
Security is worth taking seriously here given what auto browse is doing. Google says the models are trained to detect prompt injection attacks, and the system is designed to ask for user confirmation before completing sensitive tasks. That’s a reasonable baseline, but prompt injection remains an active and evolving threat when AI agents interact with arbitrary web content. This is a real risk with any browser-based agent, and Google’s framing suggests they know it.
The feature set on offer is:
- Web page summarization for long-form content
- Contextual Q&A about whatever page you’re viewing
- Integration with Google Calendar and Keep
- Image customization and generation via Nano Banana
- Auto browse for task automation (AI Pro and AI Ultra only)
So where does this leave competitors? Apple Intelligence is still mostly limited to on-device tasks and Siri improvements. Microsoft’s Copilot is embedded in Edge but has less app integration depth on mobile. Google’s position in Android gives Gemini in Chrome a distribution advantage that neither Microsoft nor Apple can easily match on that platform. This isn’t just a product update. It’s Google using Chrome as the default surface for AI, and that decision has real long-term consequences for how users interact with the web.
