How to Use Query Fanouts in AI Search

Query Fanouts in AI Search expand your core prompts into different variations that users may search across AI platforms. You can understand the queries generated for each prompt and see how frequently different variations appear across AI platforms.

You can review these variations and analyze which words are preserved, dropped, or added in the generated queries.

This helps you better understand the variations associated with your tracked queries and identify how users may phrase their searches across different AI platforms.

Query Fanouts are multiple variations of a core prompt that represent the different ways a user might search for the same information across AI platforms.

For example, if your core prompt is “dentist near me,” Query Fanouts can generate variations such as “dentist near me in New York,” “best dentist near me,” or “Hey Google, find a dentist near me.” These variations can include long-tail, voice-search, and modifier-based phrasing.

For each tracked prompt, AI Search generates multiple query variations to broaden the ways the prompt can be represented in search. Reviewing these generated queries helps you understand the different search phrasing associated with your prompts and how frequently those variations appear across AI platforms.

The Query fanouts page opens, where you can view the queries generated for your prompts and analyze them across supported AI platforms.

For each tracked prompt, AI Search generates multiple query variations to broaden the ways the prompt can be represented in search. Reviewing these generated queries helps you understand the different ways users may phrase searches related to your prompts and how frequently those variations appear across AI platforms.

Use the location dropdown to choose the locations you want to analyze. You can select All locations, a single location, or multiple locations, then click Apply to view the corresponding query data.

The Query Fanouts report shows the query variations generated for each tracked prompt. Use the report to review generated queries and understand how frequently they are used across AI platforms.

The report includes the following:

You can also use the date selector at the top of the page to view query fanout data for a specific month.

Hover over the prompt you want to review and click the three vertical dots that appear on the right. Select View details to open the detailed information for that prompt.

The Query Fanouts Per Execution section shows the number of queries generated each time the selected prompt is executed across supported AI platforms.

The chart displays query fanout activity over time for LLMs you’re tracking. Hover over a point on the chart to view the number of queries generated for each platform for that period.

Use the date-range selector to view query fanout data for a specific period, such as the last 3 months.

This view helps you understand how many query variations are being generated for a prompt and how that volume changes across AI platforms over time.

On the Query fanout details page, click the Generate Summary icon to generate an AI-powered summary of the displayed query fanout data.

The generated summary highlights key observations from the query fanout data, such as differences in query volume across AI platforms and trends within the selected time period.

You can review the generated summary to quickly understand the key insights from the query fanout data without manually analyzing the chart.

Click the three vertical dots in the top-right corner of the Query Fanouts Per Execution section to access additional actions. You can:

The Query fanout variants section shows the unique query variations generated for the selected prompt. Use this section to understand how the same prompt is phrased across different searches and how often each variation appears.

You can:

The Word Transformations section shows how words change as a prompt is transformed into query variations. You can analyze which words are added, dropped, or preserved across the generated queries.

Use the location selector to view data for all locations, a single location, or selected locations. You can also switch between the LLM tabs to compare how each AI platform transforms the prompt.

View added words

Select Added words to see the words that are introduced when a prompt is transformed into query variations.

The table shows:

View dropped words

Select Dropped words to see the words from the original prompt that are omitted in generated query variations.

The table shows:

View preserved words

Select Preserved words to see the words that remain unchanged from the original prompt in the generated query variations.

The table shows:

You can use the category selector to switch between Added words, Dropped words, and Preserved words. This lets you compare how the wording of a prompt changes across its generated query variations.