Query Fan-Out

When someone types a question into ChatGPT or Perplexity, the engine rarely searches that exact sentence. It quietly breaks the question into several more specific searches first — a comparison, a narrower version with a constraint added, a reviews-and-pricing check — and writes its answer from whatever those searches turn up. This is called query fan-out, and it matters because those hidden sub-questions, not the one the person actually typed, are what decide whether your page gets found.

Query Fan-Out models that process for any question you give it. Type a prompt, or pick one you already track, and it generates the sub-queries an AI engine would plausibly run behind the scenes, each one labelled with its type (reformulation, comparison, follow-up, and so on) and a plain-English reason an engine would run it for that specific question. This is not a capture of what any engine actually searched; no engine publishes that. It is a modelled, evidence-based estimate of the shape of the real fan-out, useful for deciding what to actually write about.

From there you can export the full set to a CSV for a content brief, or track any single sub-query as a real prompt with one click. It gets added to your tracked prompts and starts getting scanned daily across ChatGPT, Perplexity, and Gemini like everything else you follow.

It exists in two places: as a free, no-login tool anyone can try, and inside your dashboard, pre-filled with your own tracked prompts instead of generic examples — so seeing the fan-out for a question you already care about takes one click instead of a retype.