Feature
Prompt Tracking
Prompt tracking monitors the questions your buyers actually ask, tests each one against every AI search surface we track, and tells you which engines cite a competitor instead of you. The point is to track the question, not just the keyword, and to score whether this specific business can win it.
Tracked buyer questions, discovered automatically or added by you
Tested against every surface, twice a week
Named, mentioned, or absent, per engine, per question
Any weak prompt becomes a content mission in one click
The questions arrive two ways
You type them in, or discovery finds them for you. New questions arrive weekly, each one scored on two things that matter more than search volume: whether the question shows real buying intent, and whether a business your size can realistically win it. Every discovered prompt shows the keyword it came from, so you can see why it was suggested and drop it if you disagree.
You curate, and nothing is lost
Four chips sit at the top: all, discovered, custom, and dismissed. Any prompt can be set aside, and dismissal is reversible with one click from the dismissed view. A dismissed prompt stops being measured entirely, so both your attention and the measurement budget stay on the questions you care about.
Long lists stay fair
A long list is only useful if all of it gets measured. Newly added questions are picked up on the next run, and nothing sits permanently unmeasured behind the same thirty favorites. That is the difference between a hundred tracked prompts being a real number and a vanity one.
Dead questions stop spending
Questions where you hold any citation signal stay on the full cadence. Questions that keep coming back empty are checked less often, and return to the full cadence the moment a signal appears. Nothing is dropped, and you are never charged attention for a question that has gone quiet.
What it does not do
Worth knowing before you buy, not after.
- It reports which engines cite someone else on a question. It cannot make an engine cite you, and no tool can.
- A dismissed prompt is not measured at all. That is the intent, but it does mean a dismissed question will not quietly resurface with new data.
- Per-prompt sentiment correlation is not built. What ships today is axis-level sentiment on its own tab.
Questions
What buyers ask before they start
Can I add my own prompts?
Yes. Add any question with the custom prompt control and it enters the same testing cadence as a discovered one. You can also dismiss anything you do not want measured, and restore it later.
How often are prompts tested?
Twice a week against every surface we track, asking as a customer in your area. Questions that keep coming back empty are checked less often and return to the full cadence as soon as a signal appears, so your measurement stays pointed at questions that are still live.
How is this different from keyword tracking?
A keyword is what someone types into a search box. A prompt is the whole question someone asks an AI assistant, and the answer that comes back either names you or does not. Prompt tracking scores the question itself, including whether this specific business can realistically win it.