Google has introduced a fundamental new concept allowing users to interact with the web. This innovative experience is made possible by a unique new suite of AI agents. With this introduction, we hope to change the way people interact with the internet. This allows AI to tackle multiple tasks on behalf of users, creating a more streamlined and tailored experience. Through the skillful rhetoric of a Powerpoint slideshow, delivered via Zoom, in a dimly-lit, pajama-clad office, that announcement came. They proposed their vision for this new ground—an agentic web.
We were excited to see Ben Thompson’s recent post on his platform Stratechery. In it, he teased out a few of the potential dangers of this agentic web. He did point out that although this idea could produce major efficiencies, it really introduces major reliability issues and user command issues. Microsoft CTO Kevin Scott has called for a “truly open agentic web.” In this decentralized web, AI agents can work throughout the internet to accomplish tasks on a user’s behalf.
This latest embrace of generative AI agents isn’t Google’s first rodeo. In 2022, the company started integrating these AI summarizes into its core search tool. This rollout was hampered with false information—known as hallucinations. These concerns made it difficult to assess whether AI technologies are prepared for scale.
Google CEO Sundar Pichai made the case for improving user engagement with the web a key focus of his company’s presentation. For example, he announced that the new AI agents are now capable of performing up to ten unique tasks concurrently. They’re going to be experiencing your web pages and content first hand. This ability is a huge departure from the search paradigms we have all known for the last two decades.
Since its founding, Google has served users a curated list of algorithmically-chosen links for any searched query. With the arrival of AI agents, we will experience a far more robust interaction with this information. The agents are capable of rich, naturalistic conversations with users. They’re able to do everything from reading web pages, summarizing their content and assisting users during ecommerce transactions.
Google has done some pretty groundbreaking work developing an Agent2Agent protocol. This protocol makes it possible for AI agents to interact fluidly with countless data sources, making their grand vision a reality. Strong technological infrastructure and connectivity will be key to empowering better collaboration between all agents. It’s the same idea as Anthropic’s MCP, a mechanism intended to increase inter-agent connectivity.
Concerns about the reliability of AI models have been emphasized recently by DeepMind CEO Demis Hassabis. He warned against the other side of this enormous potential – the ability of AI to destroy deeply entrenched knowledge. Reliability and performance assurance, he said, remains the biggest challenge.
Given all of this, it’s still very hard to believe that search in the traditional sense will look fundamentally different. Just ask our panel of experts who all think there’s still a big role to play for the status quo in the future of information access.
In response to this, Google’s new AI-powered shopping features are designed to make discovery seamless, and the experience quick with effortless check-out processes enabled. With conversational search, Liz Reid, Google’s VP of Search, stated that “what we’re opening up is a whole new era in technology.”
“We couldn’t be more excited about this chapter of Google search where you can truly ask anything […] your simplest and hardest questions, your deepest research, your personalized shopping needs.” – Liz Reid, Google’s VP of Search
The move to an agentic web would be a radical improvement in how users interact with the web content they view. Tech companies like Google are undoubtedly putting in the effort to revolutionize digital engagement. They face unique opportunities and challenges in ensuring their AI systems live up to the promise to be effective and trustworthy.
A primary facet of this evolution is recognizing human attention as a scarce commodity. A communications leader at Google noted that “human attention is the only truly finite resource,” underscoring the necessity for technologies that can navigate the vast information landscape on behalf of users.