Google has recently significantly upgraded its search and advertising systems in an effort to reshape user commercial experiences through artificial intelligence. In the updated search interface, the Gemini model can generate customized recommendation descriptions tailored to specific products while embedding chat functionality directly within ads. Users can now pose questions straight to advertisements, with the AI providing real-time responses based on merchant website data and even guiding users to fill out contact forms. Under AI-powered search modes, ad formats have become more intuitive, potentially appearing as full-screen displays or list recommendations designed to transform ads into useful conversational supplements. Meanwhile, Google is aggressively advancing its AI agent strategy. Inspired by the success of the open-source platform OpenClaw, Google unveiled a comprehensive new agent initiative at its I/O developer conference. These agents operate continuously in the background and seamlessly integrate into both Google's ecosystem—including Gmail and Drive—and external services. The flagship consumer-grade agent, Gemini Spark, supports round-the-clock cloud execution with cross-device synchronization, assisting users with tasks such as shopping coordination, schedule management, and information gathering. Additionally, Google introduced expanded capabilities via the Antigravity platform alongside the newly launched Gemini 3.5 model, which delivers substantial improvements in coding proficiency, multi-agent concurrent processing performance, operational speed, and cost efficiency compared to previous versions. With a billion-user base and deep technical expertise, Google aims to transition AI agents from laboratory concepts into everyday tools. Although currently lagging slightly behind competitors in this domain, Google leverages its massive scale and ecosystem advantages during what represents a critical phase in determining practical utility for AI agents. Success could position Google as the industry leader, whereas failure might undermine broader prospects for monetizing AI agents commercially.