Query fan out tool: One question in. Every search behind the answer.

The engine reports them. The sub-queries arrive as a field on the same response that carries the answer text and the citations, and Verand stores the array exactly as it came back. Nothing is modelled, reverse-engineered from the citations, or written by another AI. That is why a response carrying none shows an empty state rather than a guess.

Mostly ChatGPT and Claude, Gemini only occasionally. The other engines rarely or never report the searches behind their answers, so your fan-out volume follows which engines your plan covers.

Because they are conversational phrasings, not terms people type into a search box. Get keyword metrics pulls volume, difficulty and intent on demand, and when a query has none the card says so in those words rather than showing a zero. Volume tends to appear only once several fan-out queries cluster into one head topic.

In the prompt and response drawers the list dedupes on exact text, lowercased, and nothing else. “accredited investors Texas Georgia” and “accredited passive investors Texas Georgia” stay two rows because they are two phrasings the engines really used. Collapsing them would hide the thing you came for, which is the wording. In Opportunities they become one card, because one article would answer both, and the card still quotes word for word the search it starts from.

It adds one active prompt, and prompts are pooled across every site on your account rather than granted per site. If the pool is full the add is refused and tells you so; pausing a prompt anywhere frees the slot at once. There are no overage charges on any plan.

Yes. Every routine and every check is on every plan, and fan-out is captured, shown in the prompt drawer and the response drawer, and listed in Opportunities on all of them. Plans differ by volume, sites and how many assistants you are scanned against, and that last one is what decides how much fan-out you see.