Query Fanout as an Entity Optimization Strategy for AI Search

I’ve written before about how entity optimization for AI search isn’t a checklist of tactics but rather a holistic process that involves reducing ambiguity and increasing retrievability for your brand’s key entities.

One method you can use to uncover entity relationships to optimize toward is to analyze query fanout data.

Fanout queries reveal more than target keywords. They can expose the concepts, entities, and relationships surrounding a prompt.

For example, say you want visibility for the prompt: “best AI visibility tools for enterprise SEO teams.”

A keyword-oriented strategy might focus on “AI visibility tools,” “enterprise SEO tools,” or “AI search tracking.”

But analyzing the query fanouts may reveal concepts like “ChatGPT citation tracking,” “enterprise SEO reporting,” or “AI brand monitoring.”

Now you can see that the AI assistant isn’t dealing with one topic. It is potentially traversing a network of concepts and entities.

That creates your entity optimization opportunity.

In this article, I’ll introduce a practical method for extracting and analyzing fanout queries for entity optimization purposes using an AI assistant.

Here’s a video summary, as well:

Step 1: Export Your Query Fanouts

You can find these manually by inspecting the ChatGPT conversation and looking in the Network tab:

The problem with this approach, besides being cumbersome, is that it changes over time as ChatGPT updates.

Another option is to use a Chrome extension, but these can be unreliable (as in, their fanout capabilities can break when ChatGPT does an update, as mentioned above):

Still another option is to use your prompt-tracking tool of choice. Most of the major ones like Profound and Peec AI have a fanouts section:

The problem here is the fanouts are grouped by topic, when I was hoping to view them at a prompt-level.

That is to say, all of the approaches are imperfect, so I’d recommend choosing the least imperfect one that you can do in the quickest amount of time with the surest result.

For this article, I’ll filter my prompt-tracking tool by the topic “Generative Engine Optimization” and export those fanouts.

Step 2: Use an AI Assistant to Extract the Main Entities and Their Relationships

That second part is important: we’re not merely interested in the entities themselves but how they relate to one another. This information gives us more ammunition for content topic ideation.

Upload your exported fanout queries into your AI assistant and prompt it to extract the main entities and their relationships.

Here you can see a list of entity relationships (albeit based on cooccurrences) in a table extracted from a ChatGPT response below:

Entity relationship	Queries connecting them
GEO ↔ AI search	73
GEO ↔ methodology/research	40
GEO ↔ Google AI Overviews	37
GEO ↔ content optimization/discovery	28
Google AI Overviews ↔ organic traffic/clicks	28
Google AI Overviews ↔ methodology/research	26
Bing/Microsoft ↔ Copilot	22
AI search ↔ citations/sources	22
GEO ↔ organic traffic/clicks	20
GEO ↔ citations/sources	19
Google AI Overviews ↔ AI Mode	17
GEO ↔ original Princeton GEO research	17
Google AI Overviews ↔ Google Search Central	16
Perplexity ↔ citations/sources	16
ChatGPT/OpenAI ↔ publishers	16
ChatGPT/OpenAI ↔ citations/sources	15
ChatGPT/OpenAI ↔ OAI-SearchBot	13

That gives me the foundation to start creating content topics that cover those relationships.

Note: Be careful calling everything an entity or a relationship. AI assistants are limited in their ability to do this work, so you’ll need to apply your own knowledge as well. Also, don’t base topical priority around rate of occurrence. Prioritize topics that align with your expertise, are commercially useful, and are underrepresented in your current content.

Step 3: Build an Entity-Led Content Calendar

Within my tracked GEO topic, the AI assistant identifies a range of recurring entity (and concept) associations across the fanout queries.

Possible content topics might include the relationship between GEO and content optimization, GEO and organic traffic, or ChatGPT and publishers.

I can make up such topics myself, or ask the AI assistant to formulate suggestions based on the entity relationships it uncovered in the fanout queries:

Suggested topic	Entity relationship it covers	Core angle
Entity Optimization for AI Search: A Practical Framework	Entity optimization ↔ AI search ↔ GEO	Define what entity optimization means specifically for generative search.
How Knowledge Graphs Influence AI Search Visibility	Entities ↔ knowledge graphs ↔ AI search	Explain how entity relationships help machines understand brands, people, products, and topics.
Entity SEO vs. GEO: What’s Actually Different?	Entity SEO ↔ GEO ↔ AI search	Show where traditional entity SEO overlaps with and differs from GEO.
Structured Data vs. Entity Optimization: What’s the Difference?	Structured data ↔ entities ↔ knowledge graphs	Make the case that schema is evidence about an entity, not the entity strategy itself.
How AI Search Engines Decide Which Sources to Cite	AI search ↔ sources ↔ citations	Break down retrieval, source selection, authority, relevance, and citation.
From Crawled to Cited: The AI Search Content Pipeline	Crawling ↔ retrieval ↔ citation	Introduce stages like discoverable → understandable → retrievable → selectable → citable.
How Brand Mentions Build Entity Authority for AI Search	Brand mentions ↔ entity authority ↔ GEO	Explore whether unlinked mentions and consistent associations strengthen machine understanding.
Brand Positioning as an Entity Optimization Strategy	Positioning ↔ entity identity ↔ GEO	Explain why clearly owning a topic/category may make a brand easier for AI systems to characterize.
How to Build Strong Brand-Topic Associations for AI Search	Brand ↔ topic/entity ↔ authority	Practical tactics for establishing what a brand should be known for.
What Makes a Brand an Authoritative Entity?	Entity ↔ authority ↔ citations	Separate traditional domain authority concepts from broader entity authority.

I can also compare those relationships between entities against my existing content to determine if I have optimization opportunities.

Note: The AI assistant may tend to drift in its content ideation away from the fanouts into general topics. That defeats the purpose. Focus on the topics that relate to your fanout queries. In my case, good topics might include:

• Does GEO Actually Increase Organic Traffic?
• What the Original Princeton GEO Research Really Found
• How Google AI Overviews Affect Organic Clicks
• How ChatGPT Selects and Cites Publisher Content
• GEO for Google AI Overviews vs. AI Mode
• How Perplexity Uses Sources and Citations
• What OAI-SearchBot Means for ChatGPT Visibility

Step 4: Measure AI Visibility and Revisit the Fanout Queries

As I begin creating content informed by the fanout queries, I’d keep an eye out to see if my owned content was being retrieved more often as sources for AI-generated answers. Such data would be visible in my prompt-tracking AI visibility score, for example. There would be a lot of variables at play, so you’re looking for directional evidence.

We also know that fanout queries are probabilistic, like AI responses themselves, and so we’d want to keep an eye on how the fanouts evolve so that we can understand any evolutions in entity mentions or relationships that need content coverage.

Takeaway

Fanout queries can reveal entity relationships to optimize toward as part of a broader GEO strategy.

The fanouts include entities and their relationships, which we analyze with an AI assistant, resulting in content opportunities that can ultimately lead to stronger associations.

That’s the cycle of entity optimization.

The goal isn’t to create pages about every entity that appears in a fanout query, but rather it’s to understand which concepts and entities AI search repeatedly connects around your target prompts, then strengthen the relevant relationships your brand should belong to.

If you need help with entity optimization for AI search, contact me. I’m an independent SEO/GEO consultant helping brands and agencies.

Until next time, enjoy the vibes:

Thanks for reading. Happy optimizing!

Related posts