See what AI engines really do with your query, in 30 seconds, no signup required. When buyers ask ChatGPT, Perplexity, Gemini, or Claude about your category, the engine runs five to thirty searches behind the scenes before answering. This tool shows you that hidden work.
Last updated: June 2026. Built by Astiva AI, the Competitive Intelligence platform for AI Search and Visibility. Brands compete on recommendations, not rankings.
Definition: Query Fan-Out. Query fan-out is the retrieval architecture by which an AI search engine decomposes one user query into multiple synthetic sub-queries, retrieves passages for each in parallel, and synthesizes the results into a single response naming three to five brands. The mechanism is documented in Google patent US20240289407A1, Search with Stateful Chat (August 2024), and implemented in production by ChatGPT, Claude, Google Gemini, Google AI Overviews, Google AI Mode, and Perplexity.
A single user query no longer maps to a single keyword. AI engines decompose it into a fan of synthetic sub-queries, retrieve passages for each, and synthesize one answer naming three to five brands. If your brand only ranks for the head query but misses the fan, your brand is invisible to the answer the user actually sees. Query fan-out is implemented in production by ChatGPT, Claude, Google Gemini, Google AI Overviews, Google AI Mode, and Perplexity. See the full pillar guide at astiva.ai/blog/query-fanout.
Knowing the sub-query fan is the Detect phase. The Detect, Diagnose, Displace, and Prove Cycle that organizes Astiva AI follows through: Detect maps where AI engines route the fan, Diagnose surfaces which sub-queries your brand is winning or losing across the activated AI platforms, Displace generates citation-ready content closing the identified gaps, and Prove ties the resulting citation lift to GA4 sessions, conversions, and revenue from $249 a month. Full methodology, formulas, and validation cadence are published at astiva.ai/methodology.
An AI search engine breaks one user query into several synthetic sub-queries, retrieves passages for each, and synthesizes a single response naming three to five brands.
Yes. No signup or credit card is needed, with 5 fan-outs per hour per IP. Three dimensions require no account; Head-to-Head and Buyer Intent require a free account.
The free tool shows likely sub-queries. The full platform tracks which brands AI engines cite for them across 10 platforms, with gap analysis and GA4 attribution — starting at $249/month on Growth.
Yes. Running a competitor's category query reveals the sub-queries AI engines use, including buyer phrasings. Seeing which brands get cited for those sub-queries requires the full platform.
They require brand-specific inputs (competitor list and ICP definition). Both unlock with a free account.
Win AI Search.
Read the full Query Fan-Out pillar guide | Free AI Brand Visibility Analysis | Methodology | Astiva AI Pricing