Sundar Pichai, Google’s chief executive, stated Thursday that future artificial intelligence agents will access web content through a combination of human-like browsing and specialized communication protocols, particularly for e-commerce.
This dual approach is central to Google’s vision for AI integration, balancing user control with scalable, efficient data exchange.
Pichai indicated that while browser-based methods offer users direct oversight, dedicated protocols are expected to provide greater scalability but will require more time to develop, according to an interview with Alex Heath.
Google is actively supporting the Universal Commerce Protocol (UCP), an open standard designed to facilitate agent-driven commercial transactions. This protocol enables direct purchasing and seamless shopping cart transfers.
The company’s hosted AI agents, which browse the web, utilize a user agent identified as Google-Agent. These agents are classified as user-triggered fetchers and frequently bypass robots.txt directives, Roger Montti reported.
Pichai anticipated a period of “chaotic evolution” as the industry moves toward interoperable standards for AI agents, drawing parallels to the early stages of the internet’s development.
Google plans to broaden the accessibility of Gemini’s agent features to a wider user base. The company is also expanding UCP functionalities to encompass lodging and food service transactions.
The development of these protocols and browsing methods aims to shape how AI agents interact with the vast amount of information and services available online, potentially affecting traditional advertising models.
Companies like Amazon and Meta are also exploring various approaches to AI agent integration, highlighting a broader industry push toward more autonomous digital assistants.
Google believes that establishing open standards like UCP is important for fostering an interconnected ecosystem for AI-driven commerce.
Source: Search Engine Journal
