Hands-On: Gemini's Task Automation Is Slow but Impressive

Early hands-on tests with Gemini's new task automation feature reveal a technology that is simultaneously clunky and groundbreaking. Available on select Pixel and Galaxy phones, the feature allows the AI to autonomously take actions within a limited set of apps, such as booking a ride via Uber or ordering food through DoorDash. In practice, the process is slow; the AI meticulously navigates each app screen-by-screen, much like a human would, which takes considerable time. It's also currently limited in scope and can be brittle if an app's interface changes. Yet, the capability itself is profoundly impressive. It demonstrates a significant leap from an AI that answers questions to an AI that performs tasks in the digital world on your behalf. This represents the nascent stage of true agentic AI on personal devices. While far from seamless, it proves the core concept is viable. As the technology matures, it promises a future where your AI assistant can handle complex, multi-app workflo