Web Guide, AI Mode/Overviews, and the Rise of AI Search SEO: How Query Fan-Out is Reshaping Google Search
From SERP Features to AI-Curated Search
Search results pages have not been “10 blue links” for many years. With the arrival of SERP features like Knowledge Panels, Featured Snippets, People Also Ask, Top Stories, images, and video carousels, the linear results page already gave way to a more complex, multi-format environment.
Yet, even in that era, the principle of ranking → visibility → traffic still held. If you could secure a prominent position in a SERP — whether in organic listings or features — you could predictably attract clicks.
Now, with AI Overviews, AI Mode, and the experimental Web Guide, this principle is being rewritten.
Visibility in search is no longer determined only by “ranking higher” but by whether and how your content is retrieved, clustered, and summarized by AI systems.
Andrea Volpini recently described this as Google’s shift from “a web of pages” to “a web of understanding”.
Web Guide, in particular and as Andrea also says, is not just a UI test but the first glimpse of semantic SERPs powered by Gemini.
This article unpacks:
How Web Guide works compared to AI Mode and AI Overviews.
Why query fan-out is the unifying engine across all three.
The difference between navigation vs. synthesis.
What SEOs must do to adapt — a discipline I call AI Search SEO.
How Web Guide Works: Retrieval, Fan-Out, and Clustering
At first glance, Web Guide looks like a new interface: instead of a vertical list of results, you see clusters of links, each introduced by a short AI-generated paragraph.
But under the hood, the mechanics matter:
Seed query retrieval
Google first collects the ranking results for the seed query (e.g., things to do in Milan in September).
Classic SEO determines whether you’re in this first pool.
Query fan-out expansion
A custom Gemini model generates implicit sub-queries.
Web Guide retrieves the ranking results for those sub-queries.
Example fan-outs: Milan September concerts, Milan itineraries September, Milan September events 2025.
Clustering
Gemini groups the combined pool (seed + fan-out ranking results) into thematic clusters.
Each cluster corresponds to a facet of the query (e.g., “Concerts,” “Events,” “Itineraries”).
Intro generation
Gemini writes a grounded summary for each cluster, using recurring entities and facts present across multiple documents.
These intros are not synthetic answers, but paraphrases grounded in cluster sources.
Critical distinction from classic SERPs
Web Guide has no pagination. Instead, it acts like a pillar page created by Google, where clusters replace what used to be “page 1, 2, 3.”
Consequences:
If your content is not included in a cluster, you are effectively invisible — similar to AI Mode and AI Overviews.
If you are included, Web Guide can drive meaningful traffic by channeling users through its pillar-like navigation structure.
Web Guide vs. AI Mode vs. AI Overviews
Although all three features share Gemini and query fan-out, their purpose and retrieval scope differ.
Feature | AI Overviews | AI Mode | Web Guide |
|---|---|---|---|
Purpose | Provide a direct synthesized answer | Conversational exploration | Organize SERP into thematic clusters |
Core Tech | Retrieval + generative synthesis | Gemini + broad fan-out + conversational memory | Seed query rankings + fan-out query rankings → Gemini clustering + grounded summaries |
Fan-Out Style | Broad, journey-driven | Broad, iterative | Narrower, theme-driven (SERP-based) |
Grounding | Multi-doc synthesis with citations | Multi-doc synthesis in conversation | Summaries grounded in ranking documents |
Traffic Impact | High negative (replaces clicks) | High negative (absorbs attention into chat) | Moderate/uncertain → preserves click model but shifts visibility |
Best SEO Play | Provide unique facts to be cited | Cover breadth of intents | Rank for seed + fan-outs, write entity-rich intros |
Annotated Process Flow: Web Guide vs. AI Mode
To fully grasp the differences between Web Guide and AI Mode, it helps to look at the process step by step.
Both rely on Gemini and query fan-out, but they diverge in how documents are pooled, clustered, and surfaced.
The table below breaks down each stage of the pipeline — from query input to user interaction — and highlights the distinct SEO levers you need to pull to succeed in each environment.
Step | Web Guide | AI Mode | SEO Lever - Web Guide | SEO Lever - AI Mode |
|---|---|---|---|---|
1. Query Input | User query | Same | Optimize title, H1, intro for seed query aka classic on page SEO | Same |
2. Initial Retrieval | Collects ranking docs for seed query | Same | Rank for seed query via topical authority, freshness | Same |
3. Fan-Out | Generates theme-driven sub-queries | Generates broad, intent-driven fan-outs | Create pages for likely cluster sub-queries (concerts, itineraries) | Cover all intents including profiles, accessibility, weather |
4. Document Pool | Seed + fan-out rankings | All fan-out results (broader) | Optimize for both seed + sub-queries | Ensure chunks align with fan-outs |
5. Clustering | Groups docs by theme | Groups chunks by intent | One theme per page | Sections per fan-out intent |
6. Intro Generation | Grounded summary from cluster docs | Synthesis from chunks | Place entities early + month/year if needed | Entity-rich openers |
7. Output Structure | Clusters + AI-written intros + links | Unified answer + citations | Aim to be cluster intro source | Aim for multi-chunk retrieval |
8. Interaction | Users click links from clusters | Users read overview | Write click-enticing intros/meta | Ensure cited chunks link to high-value pages |
Case Study: “Things to Do in Milan in September”
Looking at the same query through the lens of AI Mode and Web Guide shows how differently Google’s systems interpret user intent. AI Mode expands broadly to anticipate diverse user journeys, while Web Guide narrows the scope into thematic clusters built from ranking results. The contrast becomes clear in the examples below.
Fan-Out vs. Cluster Probability Matrix
To make these differences more tangible, the matrix below compares AI Mode’s likely fan-out sub-queries with their probability of surfacing as distinct Web Guide clusters. It highlights where the two systems overlap, where Web Guide narrows the scope, and which intents remain exclusive to AI Mode.
AI Mode Fan-Out Sub-Query | Web Guide Cluster Probability | Reasoning |
|---|---|---|
Milan vs. Rome / Florence / Venice / Turin | Low | Comparisons rarely cluster |
Duomo / Navigli / Brera / Sforza Castle | Medium | Merged into “Guided Tours & Cultural Events” |
Outdoor activities / Weather / Clothing | Low | Folded into “General Things to Do” |
Cheap eats / Accessible activities | Low | Too niche |
Romantic / Family / Solo / Luxury | Low | Profiles don’t cluster |
Best things to do / Top 10 | High | Matches “General Things to Do” |
Free things to do | Medium | Possible if strong SERPs |
Hidden gems | Medium | May cluster if content dominates |
Events calendar | High | Matches “September 2025 Events” |
Museums | Medium | Folded into cultural cluster |
Food festivals | High | Part of events cluster |
Nightlife | Medium | May cluster if nightlife guides exist |
Day trips | High | Matches “Milan itineraries” |
How to Optimize for Both Web Guide and AI Mode/Overviews
Optimizing for Web Guide and AI Mode might look like two separate jobs. But they can coexist if you structure content intelligently.
Web Guide / Classic SEO → whole-page optimization matters (seed + fan-out rankings).
AI Mode / AI Overviews → chunk-level optimization matters (retrievability of sections).
Let’s see how for the hypothetical content we need to create or update for “Things to do in Milan in September”:
One Pillar, Many Constellations
Pillar page: target the seed query (Things to Do in Milan in September).
Subpages: build dedicated content for fan-outs (concerts, itineraries, events calendar).
Chunk-Friendly Within Pages
Use H2/H3s for sub-intents (weather, family, romantic).
Open each with entity-rich, factual sentences.
Optimize Intros Twice
For Web Guide: intros must be snippet-ready with month/year + entities.
For AI Mode: intros must also contain unique facts for citation.
Schema as Bridge
Event, FAQ, Tour schema help Gemini cluster (Web Guide) and retrieve (AI Mode).
Freshness & Temporal Splits
Web Guide: create year-specific event pages.
AI Mode: keep chunks fresh with real-time details.
Together, this ensures:
Inclusion in Web Guide clusters (Google’s pillar page).
Retrieval in AI Mode/Overviews (synthetic answers).
The New SEO Playbook: AI Search SEO
Andrea Volpini calls this Generative Engine Optimization (GEO).
Andrea and I are good friends, but I sincerely dislike the acronym GEO, and being it IMHO a specialization of SEO, I prefer to refer to it as AI Search SEO.
AI Search SEO rests on five pillars:
Pillar 1: Build for Fan-Out with Topical Authority
Anticipate likely fan-outs.
Create pillar + constellation content clusters that comprehensively cover a topic.
Pillar 2: Master Semantic SEO for AI
Use schema markup, clean topical silos, and semiotic clarity (headings, metadata, visuals, entities).
Make content AI-readable for clustering and chunk retrieval.
Pillar 3: Snippet-Like Intros
Craft factual, entity-rich, time-stamped intros.
Ensure the first 50–60 words of pages and sections are snippet-ready for Web Guide and citation-friendly for AI Mode/Overviews.
Pillar 4: Citations and Links
Citations/Mentions: essential for AI Mode and AI Overviews, where mentions across multiple sources increase retrieval probability.
Links: still critical for Web Guide, since inclusion depends on ranking for seed and fan-out queries.
Pillar 5: Branding
Strong, recognizable brands are more likely to be trusted and surfaced by AI.
Branding influences user clicks in Web Guide clusters and credibility in AI Overviews.
From Ranking to Meaning
AI Overviews are designed to absorb demand directly inside the SERP, providing the answer upfront and often reducing the incentive to click through.
AI Mode takes this one step further: by pulling users into a conversational flow, it tends to absorb even more attention and traffic, keeping people inside the generative environment rather than sending them to external sources.
Web Guide, however, works differently. Because it functions as a Google-made pillar page without pagination, it redistributes attention across curated clusters of content. This means that if your site is not selected as part of a cluster, you are effectively invisible. But if you are included, the format can become a genuine traffic generator, since it channels users directly into recommended sources.
In practical terms, you can think of it like this: AI Overviews usually reduce traffic, AI Mode reduces it even further, but Web Guide has the potential to increase it, provided you are chosen as part of its thematic clusters.
The broader picture is that the SERP is no longer a static list of links but a dynamic, AI-curated map of meaning. To succeed in this environment,
SEOs must embrace AI Search SEO: owning the themes that power Web Guide clusters, covering the journey paths that AI Mode fans out into, and providing the kinds of unique facts that get cited in AI Overviews.
Add-on
10-Step Checklist to Optimize for Web Guide and AI Mode/Overviews
Target the seed query with a strong pillar page
Example: Things to Do in Milan in September.
Ensure the seed query appears in title, H1, intro.
Anticipate fan-out sub-queries
Build supporting pages (concerts, itineraries, events calendar).
These pages increase your chance of inclusion in Web Guide clusters.
Structure content for chunk retrieval
Use clear H2/H3s aligned with likely AI Mode intents.
Open sections with entity-rich, fact-based sentences.
Write snippet-ready intros
Place month/year + entities in the first 50–60 words of pages and sections.
Write in a factual, concise style to be lifted into Web Guide cluster intros.
Include unique facts and information gain
Stats, insider details, fresh info.
These increase chances of citation in AI Overviews.
Leverage structured data (schema)
Event, FAQ, Tour, HowTo schemas clarify context for Gemini.
Improves both clustering (Web Guide) and chunk salience (AI Mode).
Design content constellations
Interlink pillar and subpages with descriptive anchors.
Signals topical authority across fan-outs.
Keep temporal content fresh
Web Guide surfaces clusters by year (2024 vs. 2025).
Maintain updated event pages with current-year references.
Make content semiotic-friendly
Use headings, bullet lists, visuals, captions, metadata.
Improves chunk interpretability by Gemini’s hybrid processing.
Track presence in AI-driven surfaces
Monitor when your pages appear in Web Guide clusters, AI Overviews, or cited in AI Mode.
Adjust structure and intros based on observed gaps.
Article by
Gianluca Fiorelli
With almost 20 years of experience in web marketing, Gianluca Fiorelli is a Strategic and International SEO Consultant who helps businesses improve their visibility and performance on organic search. Gianluca collaborated with clients from various industries and regions, such as Glassdoor, Idealista, Rastreator.com, Outsystems, Chess.com, SIXT Ride, Vegetables by Bayer, Visit California, Gamepix, James Edition and many others.
A very active member of the SEO community, Gianluca daily shares his insights and best practices on SEO, content, Search marketing strategy and the evolution of Search on social media channels such as X, Bluesky and LinkedIn and through the blog on his website: IloveSEO.net.
stay in the loop