Track the questions buyers actually ask AI about your category

Three sources feed the list. Your domain and category seed an initial set of high-intent buyer questions; your competitors contribute the questions where they already appear; and Query Fan-Out mines the answers we have captured for the follow-up questions buyers ask next. Because the third source is grounded in your real answers rather than a generic keyword file, the suggestions tend to reflect how your category is actually discussed rather than how a keyword tool files it.

Prompt discovery with keyword-volume estimates is a Professional and Enterprise feature. Where it is included, each suggestion carries an estimated demand figure so you can rank a shortlist instead of guessing. Treat those as estimates for prioritising, not as measured AI query counts: no engine publishes how often a question is asked of it, so the figure is derived from adjacent search demand. On Starter you can still add and track any prompt you like, without the estimates.

Yes, and most teams should. Suggestions are a starting point rather than a fixed list: you can add, edit, tag or remove any prompt. Questions written in your customers' own words, pulled from sales calls, support tickets or your search console, are often the highest-signal ones you will ever track. Anything you add counts against the same prompt allowance as a suggested one, so the choice is purely about relevance rather than cost.

100 on Starter, 300 on Professional, and a volume set per deal on Enterprise. The allowance counts distinct prompts, not runs: each prompt re-runs daily on every engine your plan includes, so a Professional plan across five engines produces 1,500 captured answers a day without touching the allowance. Prompts are shared across brands rather than allocated per brand, and you can retire one and replace it whenever your priorities move.

Start with the unbranded questions a buyer asks before they know you exist, because those are the ones where the recommendation is genuinely up for grabs. Then add the comparison questions that put you next to a named rival, and the factual questions where a wrong answer would cost a deal. Branded questions deserve a few slots for reputation monitoring, but they tend to confirm what you already know rather than uncover anything.

The questions where competitors appear in AI answers and you do not, on the engines your plan covers. That is a more actionable list than a raw suggestion feed, because a gap is proof the question is already being answered in your category and simply not with you in it. Each gap links back to the answer behind it, so you can see who was recommended, in what terms, and which sources the engine leaned on to get there.