You may have seen headlines about Google “reorganising” search results. Maybe you wondered if your rankings are about to mean something different. Here is the short version. Google Web Guide is a new Search Labs experiment, announced on 24 July 2025. It uses a custom version of Gemini to sort normal web results into topic groups, instead of one long list.
It is opt-in, which means people have to turn it on. Right now it only lives on the Web tab. It still shows links to websites instead of writing an answer. That makes it very different from AI Overviews and AI Mode. I think it is worth understanding before it spreads further.
Below I cover what Web Guide is and how it works. I also cover where you will see it, how it compares with Google’s other AI features, and what it could mean for your site’s visibility and traffic.
What Is Google Web Guide?
Web Guide is a version of the search results page that AI organises. Normal results rank ten or so blue links one after another. Web Guide groups pages into sections instead. Each section covers a different angle of the search.
Google introduced it in a post on its official blog, written by Austin Wu, a Group Product Manager on Search. The post calls it an experiment that helps people find information and web pages more easily. That includes pages they might not have found otherwise.
Three things define it at launch:
- It is an experiment. It sits in Search Labs, Google’s testing area for new Search features. Labs experiments can change, move into the main product or be switched off.
- It is opt-in. Nobody sees it unless they turn it on in Search Labs.
- It is about links, not answers. What you get is still a list of web pages. The AI decides how to group and arrange them.
How Google Web Guide Works
Google has shared two building blocks. One is a custom Gemini model. The other is a method called query fan-out. Neither is new on its own. But this is the first time Google has used them on the classic list of web results, not on an AI-written answer.
Gemini Reads the Search and the Pages
Google says the custom Gemini model is used to understand both the search and the content of web pages. In practice, it tries to work out the different parts of a question. Then it works out which pages answer which part.
A normal results page answers one hidden question: “which pages best match these words?” Web Guide tries to answer a bigger one: “what are the different things this person probably needs, and which pages cover each?”
Query Fan-Out Finds the Supporting Pages
Like AI Mode, Web Guide uses query fan-out. The system does not run your search just once. It runs several related searches at the same time, each aimed at a sub-topic of the first search. Then it pulls together the best results from those searches.
Google’s own help pages for its AI features describe fan-out as running many related searches across sub-topics and data sources. In practice, a page can show up for a broad search because it ranks well for one narrower sub-question. That can happen even if it would not have made the first page of the original search.
Results Are Grouped Into Themed Sections
Once the right pages are collected, Web Guide sorts them into sections. Each section focuses on one part of the search. So the reader can jump to the part they care about.
Google’s own example is a search on how to travel solo in Japan. Instead of one list, the results are grouped around themes. These include full trip guides, safety advice and first-hand travel stories. The pages are still the kind of pages you would expect to see. They are just sorted by what they help with.
Which Searches It Is Built For
Google says Web Guide helps most with two kinds of searches:
- Open-ended questions where there is no single right answer. The searcher needs to understand a topic from several angles. The Japan travel example fits here.
- Long, detailed searches written in more than one sentence. Google’s example is someone whose family lives across several time zones, looking for ways to stay in touch.
Short searches for a brand name, or simple fact lookups, gain little from grouping. I expect Web Guide to matter most for research-style searches, where people compare options and read several sources.
Where Web Guide Shows Up and How to Get It
At launch, Web Guide shows up in one place: the Web tab of Google Search. And only for people who have opted in through Search Labs.
Google added the Web tab in 2024 for people who want plain text links without AI Overviews and other features at the top. It sits next to tabs like Images, News and Videos, sometimes under “More”. If you have opted in, picking the Web tab shows the AI-organised version. Google says you can switch back to the standard Web results at any time.
Google also said it plans to bring AI-organised results to other parts of Search. That includes the main All tab, as it learns where they help most. It gave no timeline. In my view, that sentence matters more for site owners than the experiment itself. The All tab is where most searching happens.
Google’s announcement did not list which countries or languages can use Web Guide. Search Labs access depends on your account and location. So if you cannot see the experiment, it may not be offered to you yet.
How to Turn on Web Guide
If Search Labs is open to you, the steps are short:
- Sign in to your personal Google account. Search Labs experiments are not always offered on work or school accounts.
- Open Search Labs. Use the Labs flask icon in the Google app or Chrome, or go straight to the Search Labs page.
- Find the Web Guide experiment in the list and switch it on.
- Run a normal Google search, ideally an open-ended question.
- Pick the Web tab on the results page to see the grouped results.
- To go back to the normal list, use the switch-back option on the results page, or turn the experiment off in Search Labs.
I’d encourage anyone who looks after SEO to turn it on and run their own key searches. Seeing how Gemini splits your topic into sections is useful, whether or not Web Guide ever reaches the All tab.
Web Guide vs AI Overviews vs AI Mode
Google now has three AI-driven ways to show search results, and they are easy to mix up. The key difference is what the AI produces.
| Web Guide | AI Overviews | AI Mode | |
|---|---|---|---|
| What the AI produces | Web links, grouped by theme | A short written summary with supporting links | A longer written answer with links, plus follow-up questions |
| Where it appears | Web tab, for Search Labs users who opt in | Inside the main results page, when Google decides it helps | Its own space that the searcher chooses to open |
| Who decides it shows | The searcher, by opting in | Google’s systems | The searcher |
| Uses query fan-out | Yes | Google says it may | Yes |
| Status in July 2025 | New Labs experiment | Widely available | Available to all US users since May 2025 |
For site owners, the first row is the one that matters. AI Overviews and AI Mode answer the question themselves and then cite sources. Web Guide does not write an answer. It still sends people to web pages to get the answer. So the click stays at the centre.
Put simply: AI Overviews sum up, AI Mode chats and Web Guide organises.
There is overlap too. All three are built on Gemini models. Web Guide uses the same fan-out idea as AI Mode. If you have been thinking about how your content does when Google breaks a question into sub-questions, that work helps with Web Guide as well.
What Google Web Guide Could Mean for Your Site
Web Guide is an opt-in experiment on a side tab. It is not going to move your traffic numbers this week. But it shows a direction. A few points are worth thinking through now.
Covering a Theme May Matter More Than Position One
On a grouped page, the question is not only “where do I rank?” It is also “which section am I in, and where in that section?” A page that ranks seventh for the main search could sit at the top of a section that matches a specific need.
That works both ways. A site may rank well for the main term but only cover one angle. Searchers might then skip straight to a different section that better fits what they want.
Pages That Answer a Clear Sub-Question Could Gain
Fan-out runs searches for sub-topics. So a focused page that answers one part of a broad question well has a way into results it might not otherwise reach.
For example, a small travel site with a strong page on Japanese rail passes might struggle to rank for “how to travel solo in Japan”. Under Web Guide, that page could appear in a section about getting around. That is because it is a good result for that piece of the question.
Clicks Stay Part of the Experience
Many publishers worry that AI Overviews and AI Mode give people what they need before they click. Web Guide works in a different way. Every item in it is a link to a website. No AI-written answer replaces the page.
Whether that means more or fewer clicks for any one site is unknown. Google has not shared data on Web Guide clicks. I would be careful with any claims about its traffic effect until there is real evidence.
It Will Be Hard to Measure on Its Own
Google has not said how Web Guide will show in Search Console. It is an opt-in experiment on the Web tab, so the numbers are likely too small to spot in your reports anyway.
Keep watching your Performance report as normal. Be careful about blaming or crediting Web Guide for changes. In July 2025 there are bigger factors at play, including the June 2025 core update, which has just finished rolling out.
How to Prepare Your Content for Web Guide
There is no special markup, file or setting for Web Guide. Google has not published separate advice for it. For its AI features in general, Google says a page needs to be indexed and allowed to show with a snippet (the short text under a result). It also says normal SEO best practice applies. My approach is simple: make what already works stronger.
Cover the Sub-Questions People Really Have
Take your most important topics and list the separate needs behind them. For a search like “best accounting software for small business”, that might include pricing, ease of use, payroll features, integrations and moving from spreadsheets.
Then check if you have a clear, useful answer to each. That might be one full guide with well-labelled sections. Or it might be a hub page that links to focused pages on each sub-topic.
Make Each Page’s Purpose Obvious
If Gemini reads pages to decide which section they belong in, a page that tries to do everything is harder to place. A page with one clear job is easier. This is the most common problem I see on content-heavy sites.
- Write a title that says exactly what the page covers. The ideas in our guide to writing title tags Google keeps apply here too.
- Use clear H2 and H3 headings that match the questions people ask, not clever labels.
- Answer the main question near the top, then add depth.
Offer Something the Other Results Do Not
Google’s Japan example included a group for personal travel stories. That hints at something useful. Grouping by type of answer makes room for content that is different in kind, not just a copy of the top result.
First-hand testing, original photos, real pricing, worked examples and honest downsides set a page apart from ten near-identical summaries. They are also what Google’s helpful content advice has asked for all along.
Link Related Pages Together
If you cover a topic across several pages, connect them with clear internal links and descriptive link text. This helps Google see how your pages relate. It also helps a reader who lands on one part of your topic find the rest.
Test Your Own Key Searches
Once Web Guide is on, run a handful of your most valuable searches and note:
- Which sections Gemini creates for the topic.
- Which section, if any, your pages appear in.
- Which competitors appear in the sections where you do not.
- Whether a section has no strong result, which could be a content opportunity.
Even if Web Guide stays an experiment, this gives you a quick view of how Google’s AI sees the shape of your topic. I find that very useful for content planning.
What We Do Not Know Yet
I want to be clear about the gaps, because a lot of early comment will fill them with guesses.
- Rollout timing. Google said it wants to bring AI-organised results to the All tab, but gave no date.
- Availability. The announcement did not list countries or languages.
- Reporting. There is no Web Guide data in Search Console. Google has not said if there will be.
- How sections are ordered. Google has not explained how it decides which groups come first, or how pages are ranked inside a group.
- Whether it lasts. Search Labs experiments are tests. Some move into Search, some change shape and some are dropped.
Web Guide is also part of a wider change in how people reach websites. AI tools more and more sit between searchers and pages. That goes from Google’s own features to assistants like the one in our piece on ChatGPT agent browsing websites. The common thread: clear, well-structured, truly useful pages are easier for any system to understand and recommend.
The Bottom Line on Google Web Guide
Google Web Guide is an opt-in Search Labs experiment. It uses Gemini and query fan-out to group web results by theme on the Web tab. Unlike AI Overviews and AI Mode, it organises links instead of writing answers.
It will not change your traffic overnight. But Google’s plan to test it in the All tab makes it worth watching. Turn it on, run your key searches and see which sections your pages land in.
Then use what you learn. Fill the sub-topics you are missing, make each page’s purpose obvious and give searchers something the other results do not. You can read Google’s own description of the experiment in its Web Guide announcement. Its general view on AI features is in AI features and your website.
