Where search is heading
Google I/O 2026 offered perhaps the clearest view yet of where search is heading. Across the company’s announcements, the underlying direction was consistent: users will increasingly delegate discovery, comparison and transaction tasks to AI systems.
Rather than repeatedly searching for information, users may instruct an AI agent to monitor developments on their behalf. Instead of comparing products across multiple websites, AI systems may aggregate information and make recommendations directly. Booking a restaurant, purchasing a product or planning a trip could become tasks completed by software rather than by the user navigating between websites.
The Universal Cart and what it signals
One of the clearest examples of this shift is Universal Cart, Google’s proposed shopping layer that connects Search, YouTube, Gmail and Gemini. The idea is straightforward: users can discover products across Google’s ecosystem while AI systems manage comparison, recommendations and purchasing decisions. In this model, the shopping journey becomes less about visiting individual websites and more about ensuring products are visible, understandable and actionable within Google’s own environment.
For businesses, the implications are significant. Visibility may become less dependent on rankings and more dependent on whether AI systems can access, interpret and trust their information. Structured data, machine-readable content and accurate product feeds become increasingly important when algorithms are making decisions on behalf of users.
The contradiction emerging in real time
Yet there is an interesting contradiction emerging at precisely the moment this vision is being articulated. In May, alternative search engines experienced increased attention following Google’s latest AI announcements. DuckDuckGo reported a significant rise in app installs, while interest also grew in search experiences that minimise AI-generated responses. Ecosia, meanwhile, continues to attract users through a proposition centred on privacy, environmental impact and a search experience distinct from the increasingly AI-driven direction of larger platforms.
These trends suggest that the future of search may not be converging around a single model. While technology companies are investing heavily in agentic systems, some users appear to be actively seeking alternatives that preserve a more traditional relationship with the web.
“There remains a meaningful distinction between using AI as an assistant and allowing AI to become the primary interface between users and information.”
Not a rejection of AI — a question of control
This is not necessarily a rejection of AI itself. Many users are likely to embrace AI tools where they remove friction from complex tasks. The appeal of having software compare prices, monitor information or complete bookings is obvious. But there remains a meaningful distinction between using AI as an assistant and allowing AI to become the primary interface between users and information.
The rapid growth of AI-powered search has often been discussed as though it represents a straightforward replacement for traditional search behaviour. The evidence emerging from user behaviour suggests something more nuanced. As AI becomes more deeply embedded within major search platforms, some users are choosing products that offer greater transparency, fewer generated answers and more direct access to sources.
A fragmenting landscape, not a wholesale transition
The result may be a fragmented search landscape rather than a wholesale transition from one model to another. One segment of users will increasingly rely on AI systems to filter information, evaluate options and complete tasks. Another will continue to value direct exploration, source visibility and greater control over how information is discovered. Both preferences are rational responses to an increasingly complex digital environment.
What this means for publishers and marketers
For publishers, marketers and website owners, this creates a dual challenge. Content must become accessible to AI systems that are increasingly influencing discovery and decision-making. At the same time, it must continue to serve audiences who actively seek unmediated access to information.
The significance of recent search trends is therefore not that AI is winning or losing. It is that the market appears to be splitting — and strategies built around a single future model carry real risk.
The takeaway
The question is no longer whether AI will reshape search — it already is. The more useful question for brands is whether their content strategy is built for both the users who want AI to decide for them and the users who want to decide for themselves. Serving only one is increasingly a choice to be invisible to the other.
