Fan-Out Technique: More Than “Optimizing for AI”

LLM-based platforms generate AI answers by running several queries across their training data. If they don’t have enough confidence in an answer, they will also use those multiple queries to perform multiple searches.

Then they consolidate all the findings from these searches into an answer, which not only answers the initial prompt but also includes additional information to help a user more in solving their problems.

Google refers to these concurrent queries as “fan-out” queries but a similar (not yet officially named) technique is used by ChatGPT.

SEOs quickly jumped into starting new “fan-out” optimization services and writing guidelines on how to do it. But I think one important aspect is being missed in the clutter: It is much more than optimizing for LLMs. It is understanding (end engineering for) buying journeys.

But first, let’s figure out how to know where AI platforms may be fanning their research out:

How to Extract ChatGPT’s “Fan Out” Searches

Knowing which fan-out queries AI-driven platforms are using to generate answers helps publishers and business owners get a better understanding of how AI understands usefulness and searching journeys.

There are a few new tools allowing you to easily see and export ChatGPT fan-out queries when you prompt it to research anything. These tools only work when ChatGPT is searching the web to compile an answer, and they only work in the Google Chrome browser.

This free Google Chrome bookmarklet allows you to access all search queries ChatGPT uses when finding an answer to your prompt. It also allows you to see all of ChatGPT’s reasoning behind searching for those queries. For example, for the [best organic skincare brands to try] prompt, ChatGPT’s fan-out searches were:

• best organic skincare brands 2025
• USDA certified organic skincare brands list
• emerging organic skincare brands 2024

And the reasoning contained some very enlightening logic, which would definitely help content writers as well as the product development team:

I’m looking up the best organic skincare brands and will include a variety of options based on reviews from trusted sources. I'll focus on a mix of established and emerging brands, including certifications, hero products, and price points. I’ll make sure to reference several sources like Allure and Healthline, and avoid any inaccurate brand claims. I'll also provide helpful tips on evaluating organic certifications, like USDA Organic. Not sure if images are required, but I’ll skip that for now since the user didn’t ask. Time to dive into the search!

From this login, there are a few cool observations on how ChatGPT is doing its research:

• It wants to diversify: It won’t just mention best-known options, but it is willing to include a good mix of established and new brands
• It is willing to focus on trusted sources. We don’t quite know how it establishes trustworthiness, but I guess, just like humans, it knows the trusted brands. This is similar to Google’s Knowledge Graph: When you are treated as a trusted entity, this is hard to lose.
• It wants to provide additional information, like how to avoid inaccurate brand claims and how to evaluate organic certifications.

The bookmarklet lets you see and export all of this on one clean page:

To install the bookmarklet, open the Bookmarks bar in your browser (from the “View” menu), copy the code from the page, and paste it into the Bookmarks bar. Any time you are having a conversation with ChatGPT, click the bookmarklet, and if ChatGPT performed any searches, the bookmarklet opens a new tab with fan-out queries and reasoning.

This Chrome extension keeps a record of all your ChatGPT searches and allows you to export all of them into a spreadsheet. It is a handy tool if you are having a long conversation with ChatGPT and want to see all its searches.

Using Gemini to Understand Possible Fan-Out Questions

Any time you are working on a keyword list, simply ask Gemini where it can be fanning out if it has to research related prompts. I prefer using Gemini for these purposes because it Google’s and runs AI Mode, so I feel like it may be providing more insight into how Google may interpret query.

And it does a great job too! I used the same keywords, so my prompts was as follows:

If someone is searching for "best organic skincare brands to try," what would be fan-out queries and questions Gemini or AI Mode would use to generate a more helpful answer

Now Gemini’s suggestions were absolutely great:

Optimizing for Fan-Out Queries

Now we are getting to the most important point of this article: Fan-out query optimization.

There are two things you need to remember before you get into strategizing:

• Understanding possible or actual fan-out queries means understanding possible buying journeys.
• At this point, we are also optimizing for AI agents as they are already running the whole buying journey and making buying decisions.

With that in mind:

Not each “fan-out” query belongs on your landing page

Some could be supporting content, some can be included on other existing pages.

I know many SEOs are tempted to simply collect all these related questions and add an FAQ on a landing page addressing all of them. The mindshift of optimizing for each instance of a query needs to go.

Yes, FAQs can be helpful and FAQ schema can increase your page’s chances to be referenced in an AI answer. But that is a short-term solution. Your real goal is to be THE ANSWER to all of these questions rather than referenced in one from time to time.

Your off-site digital context

Many of these queries should be your off-site marketing strategy: How are you positioned outside of your brand (in reviews, on Reddit, when compared to your competitors) when these questions are asked? And what should you do to be better positioned?

Strategic link building, Reddit marketing, and review management should always be part of “fan-out” optimization.

Social media strategy should also be aligned based on this research. You need to be consistently addressing these questions on your social media channels, as well as participating in relevant discussions.

Creating a consistent digital business footprint is tough but it is much more sustainable in the long run, because once you establish yourself as a relevant entity solving a variety of related problems, you are in that training data for a long time.

Your sales funnel optimization

This section of Gemini’s fan-out suggestion, for example, looks like a ready-made customer onboarding strategy:

This could be a quiz, a series of articles or both - all capturing a customer based on their specific needs and leading them to well-informed buying decisions.

To conclude:

To truly leverage the AI-driven fan-out technique, businesses should:

1. Map the extended buying journey: These insights are invaluable for understanding the various stages of a customer's decision-making process.
2. Diversify content beyond landing pages: Develop a comprehensive content strategy that includes supporting articles, blog posts, and other resources to address specific questions and provide a holistic answer to each customer’s broader problem.
3. Prioritize fff-Site digital context: Invest in strategic link building, Reddit marketing, and review management to shape your brand's perception in external conversations.
4. Integrate fan-out insights into sales funnel optimization: Design quizzes, article series, and personalized onboarding flows that capture customers based on their specific needs and guide them towards informed purchasing decisions.