Web Guide search experiment from Google

Google has unveiled an experimental new feature called Web Guide, part of its ongoing Search Labs initiatives. Designed to reshape how users engage with the search engine results page, Web Guide leverages artificial intelligence to organize links and resources more intelligently. This experiment aims to help people discover information on the web more efficiently by grouping relevant links and enhancing the context around search queries.

Web Guide search experiment from Google organizes results with AI

At its core, Web Guide uses a customized version of Gemini, Google's powerful AI model, to analyze and organize search content. This system not only interprets the user’s query but also evaluates the broader context of web pages across the internet. The result is a search experience that surfaces information in a more structured and meaningful way than the traditional results page.

Rather than presenting a simple list of links, Web Guide clusters them into categories related to different aspects of a user's search. For example, someone looking up travel tips for Japan might see results grouped into areas like transportation, cultural etiquette, accommodation suggestions, and food recommendations. This categorization helps users quickly navigate to the specific slice of information they’re most interested in.

The new system builds on techniques also used in Google’s AI Mode, notably the query fan-out technique. This approach involves issuing multiple, related searches simultaneously, each with a slightly different phrasing or angle. By doing so, Web Guide increases its chances of identifying the most relevant and comprehensive results from across the web.

For instance, a complex or multi-sentence search like, “My family is spread across multiple time zones. What are the best tools for staying connected and maintaining close relationships despite the distance?” would typically be difficult to answer with a single page of results. Web Guide, however, can fan out the query into sub-questions such as “best video calling apps,” “group scheduling tools,” and “ways to stay emotionally connected,” organizing the answers accordingly.

This structure not only streamlines the search process but also improves discoverability. Users may uncover resources and websites they wouldn’t have otherwise seen through conventional search methods.

Initially, Web Guide is available to Search Labs users who opt in. Once activated, participants can find the new feature within the Web tab of Google Search. The Web tab itself remains fully accessible, and users can switch back to the standard view at any time, allowing for flexibility and comparison.

While the initial launch focuses on the Web tab experience, Google has stated that it intends to expand the feature's presence. As the experiment evolves, AI-organized results may begin to appear in additional parts of Search, including the “All” results tab, where a broader blend of sources and formats is usually displayed. This gradual rollout allows the company to study how AI structuring affects user behavior and which situations benefit most from the new layout.

Web Guide is especially helpful for open-ended and exploratory searches, situations where a user might not know exactly what they’re looking for or how to phrase it. Instead of forcing users to iterate and refine their own search terms manually, the system helps by identifying and offering alternate perspectives automatically.

For example, someone who enters a vague query like “how to solo travel in Japan” might receive segmented results on language tools, safety tips, budget planning, cultural norms, and more. This mirrors the kind of nuanced advice one might get from speaking to multiple travel experts, all delivered from a single search input.

Similarly, detailed multi-part queries benefit from AI’s ability to dissect and respond to layered needs. The feature reduces friction for users who would otherwise need to perform several separate searches to answer a single complex question.

It’s important to note that Web Guide is still experimental and part of the broader Search Labs program, which allows Google to test early-stage ideas with real users in the field. Feedback from participants will help shape the final direction of the feature, including whether and how it might be integrated more permanently into Google’s mainstream search interface.

By offering AI-driven structure to an otherwise linear list of links, Web Guide continues Google’s long-standing efforts to improve the utility and discoverability of the open web. Rather than replacing web results with AI-generated content, this experiment demonstrates a hybrid approach, enhancing traditional search with modern AI techniques while keeping the focus on the web itself.

• Web Guide is an AI-powered feature within Search Labs that reorganizes web results by topic.
• It uses a custom Gemini model to understand both search queries and online content deeply.
• The feature clusters links by subtopics, improving the visibility of relevant information.
• Web Guide operates using query fan-out, submitting multiple variations of a query to retrieve broader and more useful results.
• Initially available in the Web tab for opted-in users, it may expand to the All tab in the future.
• Ideal for exploratory searches or multi-sentence, detailed queries.
• Users can easily switch between Web Guide and the standard results view.