Know exactly which AI questions you win, and which you lose

Yes. Every prompt can carry tags for product line, funnel stage, region or campaign, and the rankings table filters by any of them. There is also a visibility-by-tag view, so you can see that your comparison questions score well while your problem-aware questions do not, instead of averaging both into one number. Tags are included on every plan, and a prompt can hold several, so the same question can sit in more than one segment.

Yes. Every row opens the full answer that engine returned for that prompt on that date. That is usually where the real insight sits: not that you placed fourth, but that the answer described you with a discontinued product name, or recommended a rival for a use case you serve better. Position and sentiment are both derived from the same stored text, so the score and the passage behind it never disagree.

Every result is stored with its timestamp from the day tracking starts, and history is kept for as long as your account is active, so position over time is a continuous series rather than a rolling window. There is no back-fill for the period before you started, because an engine's past answers cannot be re-run. The practical consequence is that it is worth switching tracking on before a launch or a migration rather than after one.

Being named is the floor, not the goal. A result reads as a win when you are mentioned, sit near the front of the answer, and are described in terms that would help rather than hurt a buyer, so position and sentiment both feed the outcome. The table keeps mention status, position and sentiment as separate columns for that reason: a first-place mention wrapped in a caveat about your pricing is a problem to fix, not a result to celebrate.

There is no single ranked list to read. Each engine answers each question in its own words, so the unit here is prompt by engine, and the same question can put you first on ChatGPT and leave you out entirely on Perplexity. Rankings shows that grid rather than a blended average, because an average hides precisely the engine you are losing. Starter shows three engine columns, Professional five, Enterprise all seven.

Yes. Every plan exports rankings and citations as CSV, Excel or JSON, and every plan includes the hosted MCP server, so you can query the same data from Claude or Cursor. Professional and Enterprise add the REST API, the airankly CLI and a Looker Studio connector for scheduled pulls into your own reporting. Exports carry the answer text as well as the scores, so nothing useful is locked inside the dashboard.