What sub-questions does an AI answer really cover?
Enter a phrase your buyers would type. AI assistants rarely answer a question directly: they break it into related sub-questions, answer those, and combine the results, which means the sub-question you win decides the answer you appear in. This is MODELLED, not measured. Engines do not publish the sub-queries they generate, so we apply the documented shapes buyer questions take rather than pretending to read an engine's internals.
Questions
- Is this what the engine actually generated?
- No, and we will not claim otherwise. Query fan-out happens inside the model and no engine publishes it. What this shows is the families of sub-question that buyer questions demonstrably fall into, applied to your phrase. Treat it as a coverage checklist, not a measurement. Everything on this site that IS measured says so and prints its sample size.
- Why is the output the same every time?
- Because it is deterministic on purpose. If we asked a model to expand your phrase, you would get a different list on every run, and a tool whose answer changes when nothing changed teaches you to distrust it. The same phrase gives the same list, so you can compare two phrases and know the difference came from the phrase.
- How do I use this?
- Read down the intents and ask, honestly, whether you have a page that answers each one well. Comparison, pricing and trust questions are answered from third-party sources far more often than vendors expect, which is usually where the gap is. Then check whether engines cite you on those questions rather than guessing.
- Can you show which of these I actually get cited for?
- Yes, but not here. That needs real sampling against your prompt set, which is what the paid measurement does: it asks engines real questions and records which sources they return, with the sample size and confidence interval printed beside every figure. This free tool is the map; the measurement is the territory.
Check something else
- How many people search for this, roughly?Returns search-volume estimates with the vendor named on the result and unknown terms shown as no data, never as zero searches.
- Where does your site actually rank for a search?Reads the live Google results for one term and reports your organic position, printing the top 10 so the number is auditable rather than asserted.
- Which of your pages should link to each other?Suggests internal links from your sitemap AND shows the matched terms and score behind each one, so you can disagree with the reasoning.