Google AI Mode: Query Fan-Out Explained
Source: digiday.com
Understanding Google's AI Mode and Query Fan-Out
Google’s AI-powered search, AI Mode, uses a technique called “query fan-out” to answer questions, differing from traditional search. AI Mode, which recently launched in the U.S., uses Google’s Gemini to analyze the web and provide summarized answers. Publishers are evaluating its impact on SEO.
The way AI Mode obtains information involves more behind-the-scenes searches. This makes it difficult for publishers to optimize their content to appear in those search results. SEO experts are still testing the feature.
How Query Fan-Out Works
When a user asks a question in AI Mode, the AI model breaks it down into multiple related search queries. According to Mike King, query fan-out considers the “subintents” behind a search query. For example, “best sneakers for walking” could be broken down into “best sneakers for men,” “best sneakers for walking in different seasons,” “sneakers for walking on a trail,” and “best slip-on sneakers.”
This process happens in real time. The AI model gathers information and generates a synthesized answer. Adithya Hemanth said that the AI model predicts useful information, even if the user didn't ask for it directly, and includes it in a direct answer.
Query fan-out is the foundation of AI Overviews and helps AI Mode answer complex queries like “book a vacation for a family of 5 around the U.S.” This could involve searches for “family-friendly activities,” “road trip for families,” and “family of 5 travel ideas.” Lily Ray explained that query fan-out anticipates the user's next steps to keep them within the search experience.
Implications for SEO
Traditional search relies on keywords, where one query leads to one set of results. However, with AI Mode, one query leads to multiple queries and result sets, according to Olaf Kopp. Traditional search analyzes an entire webpage for results, but query fan-out uses relevant passages from pages in its answers, he added.
Google does not reveal the queries generated by query fan-out. Mollie Ellerton noted that while the concept isn't new, SEOs are now dealing with the unknown and optimizing for it.
In AI Mode, the goal is no longer ranking for a single keyword. SEOs told Digiday that it's now about providing useful information across many sub-queries, which means fewer clicks to publisher sites. Ray noted that it is hard to get an external click and the feature seems designed to discourage external linking.
Hemanth believes some users will still click through for more information, but AI Mode responses will likely capture most of the intent. The SEO community expects search to send less referral traffic to publishers’ sites. Ellerton stated that SEO is shifting from a performance channel to brand performance.
Adapting SEO Strategies
Because Google doesn’t share data on query fan-out searches, SEOs are using Google search features and third-party tools like AlsoAsked and Profound. King said the focus is shifting to measuring visibility in AI search engines, but few tools are available. He added that SEO tools aren't prepared to support this and custom tooling is often needed.
King created Qforia, a tool that replicates query fan-out by generating related queries based on Gemini prompts. A head of SEO at a large publisher uses Qforia and is developing an internal tool. Otherwise, the primary SEO strategy remains basic, focusing on E-E-A-T, according to a publisher's SEO head who wished to remain anonymous. Hemanth said it's about doing what they’ve already been doing but in a more focused way.
Content Optimization for AI Mode
Content should be easy for AI models to understand, with clear language, structure, and bullet points. Kopp suggests passages should be two to four sentences long. SEO editorial content should focus on topics and subtopics relevant to the user's search journey. King recommends identifying keywords, checking rankings, and optimizing the passages Google is most likely to use by focusing each paragraph on a clear topic. Hemanth added that real-world experiences, such as product review videos, and original research are helpful.
King advised focusing on the customer’s specific journey instead of trying to compete for volume keywords. For example, content targeting “Denver travel” should also address queries like “Denver flights,” “how to get a hotel in Denver,” and “best places to visit in Denver,” Kopp said. Hemanth said websites need to be seen as having useful information, requiring a shift in mindset. Ellerton is using platforms like Reddit and TikTok to understand how people are searching, since Google doesn't share that data.