Query Fan-Out Generator Examples: How AI Search Engines Expand a Single Search Query

When you type a question into Google AI Mode, ChatGPT Search, or Perplexity, the AI search engine can quietly break your single query into several narrower searches, run them against its sources, and stitch the results into one response. How many searches it runs depends on the engine, the model and the question: anywhere from none or one for a simple fact to a dozen or more for a complex comparison (the measured figures are in the platform section below). This expansion is called query fan-out.

A query fan-out generator imitates this behavior: you feed it one prompt and it returns the sub-queries an AI model would plausibly send to a search engine. Most generators are simulators. They estimate what an engine might search; they do not show what it actually searched.

Google introduced AI Mode in March 2025, describing a “query fan-out” technique that runs multiple related searches at once across subtopics and data sources. On November 18, 2025, Google said Gemini 3 upgraded that fan-out process so AI Mode now runs more, and smarter, searches per question. Understanding query fan-out is essential for content teams. This article covers concrete query fan-out generator examples across six industries, plus a content strategy framework for acting on the results.

Quick start: To simulate your own fan-out queries right now (note: these tools generate plausible sub-queries; they do not expose what search engines actually fired):

• Open a free generator such as Wellows’ Query Fan-Out Generator, Position Digital’s Query Fan-Out Extractor, or Qforia (links in the tools section below)
• Paste your seed query and note every sub-query the simulator returns
• Map each returned sub-query to the eight types covered in this article to identify which content gaps you have today

What Six Fan-Out Expansions Look Like at a Glance

These sub-queries are illustrative examples written for this article, not output captured from a live engine.

Core Definition: What Is Query Fan-Out in AI Search?

Query fan out is the process by which an AI model decomposes a single user prompt into multiple narrower search queries, generates distinct queries for each angle (definitions, comparisons, pricing, risks, follow-ups), and synthesizes answers from those sources into one response.

Key distinctions:

• Query fan refers to the overall pattern of related searches an AI engine produces from one prompt.
• Query fan out generator is the specific system or tool that produces those related queries so you can analyze and act on them.
• Traditional “related searches” rely mostly on synonyms and co-occurrence data. Fan out focuses on user intent, not just lexical similarity. Intent decomposition identifies user needs and relevant attributes that traditional keyword tools miss.
• AI platforms using query fan out include Google AI Mode (launched 2025), ChatGPT Search, Perplexity, and Gemini.
• The concept is grounded in Google’s patent US11663201B2, which describes a generative model that produces query variants from a seed query, issues them separately, and combines the retrieved content into a final answer.

How a Query Fan-Out Generator Works

Every query fan out generator follows the same process. Understanding each stage helps you position content to be retrieved at the right moment.

1. Decomposition. The seed prompt is broken into intent facets. “Best project management software for remote teams” yields at least four: feature comparison, pricing, integrations, and team-size fit.
2. Expansion. Each facet produces one or more sub-queries. Running multiple angles at once ensures the AI engine explores the topic broadly rather than fixating on a single interpretation.
3. Execution. The engine runs the sub-queries, often at the same time, against its search index and other sources, retrieving passages in parallel so the final answer can draw from multiple content types at once.
4. Synthesis. Retrieved passages from each sub query’s results are merged and conflicts resolved into a cohesive response.
5. Contextual Results. The final answer is presented with inline citations. Most retrieved results never get cited (see the Writesonic figures later in this article), so passage-level quality decides which pages make it into the answer.

For content strategists: Map existing pages against each stage (decomposition through contextual results) to position passages for the execution stage rather than competing for a single keyword.

Your customer types one question. AI search engines find multiple answers from different angles. Only the best passages from those angles get cited. Your content team needs pages for each angle, not just the main keyword. Each page needs one clear, quotable answer at the top.

The 8 Sub-Query Types a Fan-Out Generator Should Produce

The eight sub-query types below are the query-variant types listed in Google’s patent US11663201B2. Each type demands a different content structure to earn the citation.

• Equivalent: same intent, different phrasing (“project management app” vs. “project management software”)
• Follow-up: the question the user asks next (“how to onboard a team to new PM software”)
• Generalization: the broader category (“business productivity tools”)
• Specification: narrower detail (“Asana vs. Monday for teams under 50 people”)
• Canonicalization: standardized or normalized versions of the query (“standardized project tracking tool”)
• Translation: cross-language variants (relevant for multilingual brands)
• Entailment: implied questions (“is Asana secure enough for enterprise?”)
• Clarification: a question put back to the user to resolve a vague term (“what does distributed mean in project management?”), which content answers with disambiguation sections

Most SEO content covers equivalent and specification sub queries well. The gap is follow-up, entailment, and clarification, types that need onboarding content, comparison and compliance content, and disambiguation sections respectively. Mapping existing pages to these eight types reveals which sub queries need new passages.

Platform-Specific Sub-Query Counts

No AI search engine publishes an official fan-out count. The figures that do exist come from different studies, models and months, so treat each as a dated snapshot rather than a fixed rule. These are the sourced numbers this article relies on:

The spread is the point: the same engine can run one search or ten depending on the model and the question. Simple factual prompts may not trigger fan-out at all, while comparative and research-style prompts trigger the most. A passage built for one engine’s retrieval may not surface in another’s.

Structure articles with standalone answer passages under descriptive headings so they can match whichever sub-queries an engine runs, and refresh your data on a regular cycle.

Example 1: B2B SaaS Query Fan Out for “project management software for remote teams”

AI search engines add modifiers like “best,” “top,” “reviews,” and year stamps when expanding B2B software queries. A query fan out generator could output: “best project management software for remote teams 2026,” “project management tools with time tracking and Slack integration,” “Asana vs Monday for distributed teams,” “ISO 27001 compliant project management software,” “how to onboard remote teams to new project management tool,” “pricing models for PM tools for remote teams,” “best project management tool for teams under 50 people.”

These group into clear clusters by sub-query type:

• Comparisons (Specification): “Asana vs Monday” style head-to-head queries are a common pattern in B2B fan outs.
• Integrations (Specification): queries about Slack, time tracking, and specific feature sets.
• Security and compliance (Specification): ISO 27001, enterprise-grade requirements.
• Onboarding and change management (Follow-up): practical implementation questions that appear after the initial purchase decision.

A SaaS company should cover these clusters (integration landing pages, feature deep-dives, security whitepapers, and onboarding playbooks) rather than targeting “project management software” as a single keyword.

Example 2: E-Commerce Query Fan Out for “best running shoes for flat feet 2026”

A generator might output these fan out queries: “best running shoes for flat feet 2026 reviews,” “stability running shoes for flat feet under 150 dollars,” “Brooks vs ASICS for flat feet runners,” “running shoes for overpronation with wide toe box,” “orthotic-friendly running shoes 2026,” “running injuries common with flat feet,” “flat feet running shoes women vs men differences.”

Each group of fan out queries generated reflects a different facet of user intent, mapped to sub-query types:

• Price sensitivity (Specification): “under 150 dollars” captures budget-conscious shoppers.
• Brand comparisons (Specification): “Brooks vs ASICS” shows how AI search engines produce head-to-head queries automatically.
• Medical and safety concerns (Entailment): “running injuries common with flat feet” shows how fan out adds risk queries the user never typed.
• Maintenance and lifecycle (Follow-up): follow-up questions the user hasn’t explicitly asked yet.

An e-commerce brand can cover this query fan with price-tiered buying guides, brand comparison pages, injury-prevention FAQs, and maintenance articles, ensuring the brand appears across multiple fan out queries rather than just the original query.

Example 3: Local Service Business Query Fan Out for “emergency plumber near me”

Local and “near me” queries trigger location- and time-aware fan out behavior. Fan out patterns here are distinctly different from informational or commercial queries because urgency, credentials, and geography dominate.

A generator might produce: “24 hour emergency plumber in Brooklyn,” “plumber for burst pipe response time under 1 hour,” “average cost of emergency plumbing visit at night,” “licensed plumber reviews Brooklyn,” “what to do before emergency plumber arrives burst pipe,” “plumber who accepts insurance for water damage.”

Key fan out patterns in local queries, mapped to sub-query types:

• Geographic modifiers (Specification): the AI translates “near me” into specific neighborhoods and cities based on the user’s location.
• Urgency and time windows (Specification): “open now,” “response time under 1 hour,” “24 hour”; all reflect latent intent AI systems detect automatically.
• Price range queries (Specification): users in emergencies still want cost expectations upfront.
• Safety and credentials (Entailment): “licensed,” “certified,” “reviews”: trust signals the AI model checks.

A local brand can ensure its brand appears in AI answers by publishing clear service area pages with specific neighborhoods listed, transparent pricing ranges, credentials and certifications, and emergency checklists that answer “what to do before the plumber arrives.”

Example 4: Health Information Query Fan Out for “symptoms of iron deficiency in women”

Health queries are sensitive, and AI models fan out heavily into safety, treatment, and demographic-specific sub queries. The system generates multiple sub queries to ensure the response covers clinical nuance rather than a single generic answer.

Example fan out queries: “early symptoms of iron deficiency in women over 40,” “difference between iron deficiency and anemia symptoms,” “iron deficiency during pregnancy symptoms,” “when to see a doctor for low iron levels,” “foods high in iron for vegetarian women,” “can iron deficiency cause hair loss and fatigue,” “iron supplement side effects women.”

Patterns in health fan out by type: demographic filters (Specification) for age groups and pregnancy status create distinct query paths; differential diagnosis queries (Clarification) address confusion between related conditions; a “when to see a doctor” follow-up (Follow-up) is a common pattern for health topics. Medically oriented content should cite reputable sources and structure information in small, extractable sections.

Example 5: Travel Planning Query Fan Out for “2 week itinerary japan cherry blossom season”

AI search expands broad queries into detailed planning questions and contextual variations for travel topics. The user typed one search query, but the AI needs budget data, logistics, timing, and local expertise to build a useful response.

A generator might output: “best time to see cherry blossoms in Tokyo and Kyoto 2027,” “2 week Japan cherry blossom itinerary with rail pass,” “budget for 14 days in Japan during sakura season,” “crowd-free cherry blossom spots near Kyoto,” “Japan rail pass vs regional passes for cherry blossom trip,” “Japan visa requirements for US citizens 2027,” “weather in Japan late March vs early April for cherry blossoms.”

These fan out queries cover timing (Recency), transportation (Specification), budgeting (Specification), and crowd avoidance (Follow-up). Content creators should build budget-tiered itineraries, JR Pass comparison pages, and crowd-avoidance destination guides, each addressing a distinct cluster rather than trying to cover everything in one generic article.

Example 6: Educational / How-To Query Fan Out for “learn Python for data analysis”

AI search engines and AI models break learning queries into skill milestones and resource types. The fan out here follows the learner’s journey from beginner to job-ready.

Fan out queries generated might include: “beginner Python course for data analysis 2026,” “Python libraries for data analysis pandas vs polars,” “sample data analysis projects for Python portfolio,” “learn Python for data analysis free YouTube channels,” “how long to learn Python for entry level data analyst job,” “Python vs R for data analysis 2026,” “data analysis Python certification worth it.”

These map to learning stages by type: getting started (Equivalent), tool selection (Specification: pandas vs polars, Python vs R), portfolio building (Follow-up), and career outcomes (Specification: time-to-job, certifications). Structure articles into curriculum-like topic clusters with clear headings that serve as independent answer chunks for each stage.

Query Fan vs Classic Keyword Expansion: What’s Different?

Classic keyword research produces standardized or normalized versions of the same concept: “project management tools,” “PM software,” “project management app.” These are lexical variations, not intent variations. Query fan out expansion of the same query yields: “best project management software for remote teams 2026,” “Asana vs Monday.com pricing,” “ISO 27001 project management tools,” “how to onboard a team to new PM software.” These are comparative, pricing, integration, and implementation queries with entirely different user intents.

Most fan out queries carry little or no search volume of their own, yet AI models still use them internally to ground answers. Fan out optimization therefore targets topic clusters, not single keywords.

Common Query Modifiers Fan-Out Generators Return

Beyond the eight formal sub-query types, generators consistently surface these commercial modifiers across industries. Understanding them helps you spot which buyer stage each sub-query targets:

• Comparative: “Asana vs Monday for remote teams”: head-to-head matchups; maps to the Specification type
• Best/Top lists: “best AI search engines for enterprise this year”: ranking queries; maps to Equivalent or Specification
• Price range: “affordable project management software per user”: budget constraints; maps to Specification
• Recency: “latest Python data analysis libraries”: year and freshness modifiers; maps to Equivalent with recency filter
• Location-based: “emergency plumber in Brooklyn open now”: geographic availability; maps to Specification
• How-to / Next step: “how to migrate from Trello to Monday”: procedural queries; maps to Follow-up
• Risk / Safety: “side effects of iron supplements women”: cautionary intent; maps to Entailment
• Review / Reputation: “Monday.com reviews from larger teams”: social validation; maps to Specification or Entailment

These modifiers are not separate from the eight types; they are the surface patterns the types produce in real searches. Mapping them back to the taxonomy helps you structure the right content format for each.

How to Use Query Fan-Out Generator Outputs in Content Strategy

A workflow for turning query fan out tool outputs into content plans:

1. Pick a seed topic. Choose a core topic you already rank for or want to own.
2. Run it through a query fan-out generator. Use tools like Wellows, Rankability, or Position Digital to generate a set of likely sub-queries, and run each seed more than once because the output changes between runs.
3. Cluster the resulting queries. Group related queries by intent bucket: comparisons, pricing, how-tos, safety, and so on.
4. Map clusters to pages and sections. Decide whether each cluster needs a section within an existing page, a standalone article, or a broader content hub.

Key principles: build topic hubs rather than separate pages for every synthetic fan out query; include follow up questions as subheadings and FAQs; use FAQ schema on government and health sites (Google dropped HowTo rich results in 2023 and now limits FAQ rich results to those site types) and Product schema on product pages; re-run core topics quarterly as AI search behavior evolves.

The Sub-Query-to-Content Structure Mapping Framework

The eight sub-query types map directly to specific content structures that AI engines can extract:

The sub-query types come from Google’s patent US11663201B2. The content structure for each type is this article’s editorial recommendation, not a measured citation rate.

Fan-Out Generators to Try

Most fan-out generators are simulators: they use a language model to estimate plausible sub-queries for a seed topic. A few routes show the queries an engine actually ran; those are marked below.

• Wellows: Query Fan-Out Generator. Enter a seed query; returns the likely fan-out sub-queries behind it. Free.
• Rankability: AI Search Query Fan Out tool. Groups its output into reformulations, related questions, comparisons and adjacent topics, which maps neatly onto page sections and FAQs.
• Position Digital: Query Fan-Out Extractor. Enter a prompt, choose an AI platform and a location, and it returns the fan-out queries. Free.
• Qforia (by Mike King / iPullRank). Built on Gemini, it simulates fan-out for AI Overviews and AI Mode and labels each sub-query with its type, the intent behind it and a suggested content format. Free. Simulator.
• DEJAN Query Fan-Out. Enter a prompt or a URL; it runs the prompt through a grounded model and shows the fan-out searches that model issues.
• Profound: Query Fanouts. Part of Profound’s AEO platform. It records the fan-outs from the prompts Profound runs against answer engines for its customers, so it reports observed queries rather than simulated ones.
• Gemini API grounding metadata (observed). With grounding switched on, the API response includes a webSearchQueries list of the searches Google ran for that request. Requires developer access.
• ChatGPT network requests (observed). ChatGPT’s network responses contain its fan-out queries; search the responses for search_model_queries in Chrome DevTools, or use an extension such as LLMrefs ChatGPT Search Query Extractor.
• Bing Webmaster Tools (observed). Shows a sample of the grounding queries behind your site’s citations in Microsoft’s AI answers.
• Goodie. Tracks citation metrics across AI platforms. Useful for monitoring where your domain is cited after fan-out retrieval.

Where to start by team size: Solo content marketers get the fastest value from the free tools: Wellows, Position Digital, and Qforia. Enterprise SEO teams with developer support should add the Gemini API’s grounding metadata, which returns the searches Google actually ran, and a tracker that records real fan-outs, such as Profound or Ahrefs Brand Radar.

Why Query Fan-Out Matters for SEO and AI Search Engines

Ranking first no longer guarantees visibility. You now compete across multiple fan-out queries instead of one search query. Because AI search retrieves content chunks rather than entire pages, each section of your content is a potential answer to a different sub-query.

Three-step audit to size your exposure gap: Pull your top pages by organic traffic in GA4 and add a custom channel group for AI referrers (ChatGPT, Perplexity, Gemini and similar); pages with organic traffic but no AI sessions rank but stay invisible in the AI channel. Run those pages’ primary topics through a fan-out generator and check which sub-query types your page explicitly answers with a standalone passage. Then set a baseline against the NoGood KPI targets in the measurement section below.

The NoGood Case Study: Rank #8 Beats Rank #2

NoGood tracks a SaaS company that ranks #8 for “project management software” yet appears in 67% of AI answers about project management, while a competitor ranking #2 for the same term appears in just 12%. NoGood credits the gap to topical coverage: the #8 site answers the sub-queries AI systems generate for the category (pricing comparisons, integrations, use cases, team-size recommendations), while the #2 site optimized for rankings. Before building new pages, check whether your existing high-authority pages already cover those sub-queries.

AI Citations Live in a Separate Channel from Organic Rankings

A Moz analysis of nearly 40,000 queries in the US and UK (published February 2026) found that 88% of the URLs cited in Google AI Mode were not in the top 10 organic results for the same query. In Writesonic’s April 2026 ChatGPT test, answers cited 1,257 of the 11,469 search results their searches returned, about 1 in 9; in a 30-prompt check, 84% to 87% of the domains cited by ChatGPT’s Thinking models were absent from Google’s and Bing’s top 10 for the same prompt. A Surfer study of 173,902 URLs found that pages ranking in the top 10 for both the main query and at least one fan-out query were 161% more likely to be cited in Google’s AI Overviews than pages ranking for the main query alone. Audit which pages earn AI citations via a GA4 AI-referrer channel group or Ahrefs’ AI citation tracking, then build a separate priority list for fan-out optimization.

The Business Case: Fan-Out Is the B2B Buyer’s Research Path

In G2’s March 2026 survey of 1,076 B2B software buyers, 51% said they now start their research with an AI chatbot more often than with Google, up from 29% in G2’s 2025 report. Many of those sessions trigger fan-out searches. In a Semrush survey of 519 US B2B professionals who use AI at work (March to April 2026), 92% said AI has shaped their vendor shortlist, so content that earns citations during fan-out retrieval has a better chance of reaching the shortlist before the buyer ever visits a vendor’s site.

What AI Query Patterns Actually Look Like

In Ahrefs’ study of 15 million ChatGPT fan-out queries (linked in the platform table above), the current year (“2026”) appeared in 22.9% of English fan-out queries in September 2026, “official” in 13.1%, and “price” in 8.9%. These modifiers shape the sub-queries AI engines generate during fan-out: nearly a quarter carry a recency filter, the “official” modifier favors primary-source content, and the “price” modifier generates cost sub-queries that commercial pages should anticipate.

Counter-Case: When Fan-Out Optimization Fails

Fan-out generators are directional, not absolute. Surfer’s analysis of 1,600 runs found that only about 27% of fan-out keywords stayed consistent across repeated runs, and 66% appeared only once. Some fan-out queries have no sustainable demand; prioritize recurring patterns over one-off results. Location, search history, and device type all influence which sub-queries AI engines generate for real users. Authoritas research tracking 11,203 keywords between August 2024 and January 2025 found that about 70% of the pages ranking in AI Overviews changed over a two-to-three month window. When Semrush updated four blog articles to target fan-out queries, citation counts for its tracked prompts rose from two to around five within a month, spiking as high as nine before settling, while brand mentions dipped; Semrush treats the result as directional, not guaranteed.

Designing Content Chunks That AI Models Can Easily Extract

AI systems extract specific content chunks, not entire pages, and most of what they retrieve never gets cited. Structure content so AI can lift coherent passages directly into AI answers:

• Use clear headings that match fan out query language. If a fan out query is “when to see a doctor for low iron levels,” make that a subheading.
• Keep paragraphs short (2 to 4 sentences). Each paragraph should answer one question cleanly.
• Include definitions, comparisons, and step-by-step instructions as standalone blocks, since they can be lifted into an answer without surrounding context.
• Use bullet lists for multi-point answers. For structured data: add Product schema on product pages; FAQ schema is now limited to government and health sites (Google dropped HowTo rich results in 2023 and FAQ rich results for general sites). Structured markup can still help machines read your page even when no rich result renders.
• Each section of a long article should stand on its own as an answer to one query variation.

Before-and-after example: A service area page that opens with a lengthy brand overview paragraph cannot be cited for “emergency plumber in Brooklyn response time” because no single passage answers the sub-query cleanly. Restructuring the same page to open with a concise standalone block (“Our Brooklyn plumbers respond quickly for burst pipes and gas emergencies. Service areas include Park Slope, Williamsburg, and Bed-Stuy. Available around the clock.”) creates a directly extractable answer passage for multiple fan-out sub-queries from a single query about emergency plumbing.

The Plaid Pricing Page Pattern

A study of AI pricing answers about Cloud 100 companies, run across six AI engines from 29 July to 10 August 2026 and reported by Omnius, found that Plaid’s pricing page was cited first in 42% of answers about Plaid’s pricing, against 12% for company pricing pages overall. The study’s authors credit Plaid’s information-dense billing documentation, without isolating its effect. The pattern points to detail-rich pages (pricing tables, tier breakdowns, billing documentation) outperforming general marketing pages for pricing sub-queries.

When a user asks “how much does Plaid cost,” the fan-out engine generates sub queries about pricing tiers, enterprise vs. startup plans, and competitor pricing comparisons. A pricing page with clear headings, a pipe table, and standalone answer passages for each tier answers more of these sub queries than a generic “contact sales” page. Audit your highest-value pages and restructure them with short, self-contained answer blocks.

Stripe: When Breadth and Structure Do the Work

In its query fan-out guide, Semrush presents Stripe as a mini case study: Stripe’s AI search visibility significantly outperforms competitors’ in Semrush’s AI Visibility Toolkit. Semrush points to Stripe’s solutions pages tailored to different business stages, models and use cases, each with subsections that answer subtopics directly, plus its blog, customer stories, support center and newsroom, and its frequent coverage in industry press. Semrush says the breadth and depth of on-site content and the extensive off-site presence likely played an especially important role.

A structural pattern to replicate on your highest-value pages:

• Headings phrased as tasks or outcomes (“Authenticate a payment request”), short enough to scan and specific enough to match a fan-out query
• Passages kept brief and self-contained per standalone block, with the answer first and context second
• A small number of inline links per passage, pointing to the next logical step or reference page
• At least one concrete example per section (a code snippet, a pricing row, a comparison figure)
• No filler intros: each section opens with the answer, not with context about why the section exists

Measuring Whether Your Brand Appears in AI Search Results

Traditional rank tracking for a single keyword is not enough in a fan out world. Track four signals: citations in AI answers (how often your domain is cited in Google AI Overviews, ChatGPT, and Perplexity); brand mentions across AI platforms even when users don’t mention your brand by name; coverage of fan out intent buckets (comparisons, pricing, how-tos, reviews) relative to competitors; and competitor benchmarking across multiple fan out themes.

Fan-out behavior differs by engine and even by model version (see the platform figures above), so measure each engine your target audience uses separately.

Your Fan-Out Measurement Stack

Tracking fan-out performance requires a dedicated measurement stack: Google Analytics 4 with a custom channel group for AI referral traffic, Ahrefs for AI citation tracking and competitor citation analysis, Semrush for AI visibility monitoring, Google Search Console for query-level performance data, and Bing Webmaster Tools for a sample of the grounding queries behind your citations in Microsoft’s AI answers.

NoGood suggests these KPI targets: a citation rate of 15–25% of relevant prompts in your category, a share of citations of 20% or more in core categories, and topic coverage of 80% or more of the themes your fan-out analysis identifies. Use them to set initial KPIs and track improvement over quarterly refresh cycles.

Content Freshness Drives Citation Frequency

NoGood reports that, in dynamic categories, content updated within the last 90 days gets cited 2.3 times more often than content older than 12 months. Separately, Ahrefs’ analysis of about 17 million citations found AI assistants cite URLs averaging 1,064 days old, against 1,432 days for organic results, which makes them 25.7% fresher; Google’s AI Overviews were the exception, leaning toward older pages like organic search. Regular refreshes therefore matter most for ChatGPT, Perplexity and other assistants, and less for AI Overviews.

Build a 90-day refresh calendar, prioritizing pages that currently earn AI citations and pages targeting high-fan-out queries. Refresh with real changes (new data, pricing, examples), not just a new date.

Video Resource: Query Fan-Out in Practice

In this interview from BSM, a Denver SEO company, founder Chris Raulf and Austin Willman of Digital Web Solutions explain how podcasting fits into an AI SEO and GEO strategy, a format most written content teams overlook.

What the video covers:

• Why, in the speakers’ view, YouTube is the top social source cited in Google AI Overviews, and how podcasts can use that.
• Why the transcript matters more than the audio for AI search visibility, which is what gives a recording passages that can match fan-out sub-queries.
• How one recording can be repurposed into LinkedIn posts, newsletters and other content, multiplying the pages that can answer different sub-queries.

Listen: Fan-Out Queries and AI Search Strategy

How Do You Know What Fan Out Queries Were Searched? James Dooley Interviews Sergey Lucktinov

In this 8-minute episode (June 2026), Dooley and Lucktinov walk through how large language models expand a single search into multiple background queries and how Google, Gemini, ChatGPT and Perplexity analyze, filter and select sources. What the episode covers:

• Why intent matters more than keywords once a prompt fans out into background searches.
• Fan-out limits, trust signals, listicle inclusion and brand mentions as factors in which sources get selected.
• How content gets extracted into an answer or ignored.

Put it to work: treat your content audit as a sub-query gap analysis, mapping every existing page to the eight sub-query types to find the types with zero coverage, and run your seed query through a fan-out generator before briefing writers.

Frequently Asked Questions

Q1. What is a fanout query?

A fanout query is a sub-query generated when an AI search engine decomposes a single user prompt into multiple narrower searches. When you enter one question into Google AI Mode, ChatGPT, or Perplexity, the system may create anywhere from one to a dozen or more fanout queries, depending on the engine, the model and the question, then retrieves content for each and synthesizes the results into a single answer. The fanout queries are usually invisible to the user but determine which content gets cited.

Q2. Are fan-out generators accurate?

Most fan-out generators are simulators, not interceptors. They use a language model to predict which sub-queries an AI engine would probably fire; they do not reveal what a search engine actually executed, and their output changes from run to run. Several routes show observed queries instead: the Gemini API’s grounding metadata (webSearchQueries) lists the searches Google ran for an API request; ChatGPT’s network responses contain its fan-out queries; Bing Webmaster Tools shows a sample of grounding queries for Microsoft’s AI answers; and trackers such as Profound and Ahrefs Brand Radar record fan-outs from the prompts they run. Use simulators for content planning and observed queries for verification.

Q3. Can you see Google’s real fan-out queries?

Google does not expose the fan-out queries behind AI Mode or AI Overviews in Search Console, Analytics, or any other report. The closest first-party view is the Gemini API’s grounding metadata, whose webSearchQueries field lists the searches Gemini ran for that API request. That is a developer tool, and it shows what the API model searched, not what AI Mode searched for a particular user. For non-developer teams, fan-out generators are the practical alternative: they simulate plausible sub-queries based on the engine’s known behavior.

Q4. Is there a free fan-out generator?

Yes. Wellows’ Query Fan-Out Generator, Position Digital’s Query Fan-Out Extractor, and Qforia (by Mike King / iPullRank, built on Gemini to simulate fan-out for AI Overviews and AI Mode) are all free to use. Developers who want the searches Google actually ran can use the Gemini API’s grounding metadata.

Q5. How many sub-queries should a fan-out generator return?

There is no standard number, because real engines vary widely. Writesonic logged an average of 10.5 fan-out queries per prompt for ChatGPT’s GPT-5.4 Thinking model and 7.3 for GPT-5.5 Thinking, while Ahrefs found that free, logged-out ChatGPT averaged about 1.6 in the responses that searched at all; Semrush estimates 8 to 20 or more for a complex Google AI query. Judge a generator by coverage rather than count: a useful output spans several of the eight sub-query types (equivalent, follow-up, specification, entailment and so on) and groups into a handful of clusters you can turn into sections.

Q6. How is a fan-out generator different from People Also Ask or AlsoAsked?

People Also Ask (PAA) shows questions Google has ranked for a query — it reflects historical search behavior and actual user clicks. A fan-out generator shows the sub-queries an AI engine would fire internally during retrieval — it reflects AI intent decomposition, not user click patterns. PAA is backward-looking (what users searched before); fan-out is forward-looking (what the AI will search on the user’s behalf). For AI citation strategy, fan-out output is more relevant. For traditional keyword research, PAA data is more reliable.

Q7. Is SEO dead now with AI?

SEO is not dead; it has split into two parallel channels. Traditional SEO still drives traffic for queries where users want to click through to websites. AI search is now a second channel where content competes at the passage level rather than the page level. Moz’s analysis of nearly 40,000 queries found that 88% of AI Mode citations were not in the top 10 organic results for the exact query, meaning the two channels draw on largely different pools of pages. SEO practitioners need both: traditional optimization for click-through traffic and fan-out optimization for AI citations.