What does the TofuBofu AI Visibility Index measure?

AI visibility is whether an AI engine recommends you when a buyer asks, the outcome AEO and GEO work exists to move. The TofuBofu AI Visibility Index measures it in two parts that follow the buying funnel, and we always report the pair rather than averaging it into one score: the Tofu Index (are you getting mentioned) and the Bofu Index (are you getting recommended).

A query is one buying question put to one engine, so the same buying question asked of all five engines is five queries. The Tofu Index is the share of those queries where an engine names your company: are you in the consideration set. It is a plain rate, not a hidden formula, and it is built from the same cells as the per-query and per-engine boards, so drilling into it always lands on a real query and a real answer.

The Bofu Index is the share of head-to-head comparisons, you against a named rival, where the engine picks you: are you getting recommended. It has its own denominator (rivals times engines) and comes with Fix and Dominate; the free Track scan gives you the Tofu Index.

What counts as an answer

Each engine's answer to each query is counted separately. A query that asks about your brand by name (used to check sentiment) is excluded from the Tofu Index, since asking "what do people think of Acme" and getting Acme back is not a visibility signal.

no_answer is not absence

When an engine returns nothing usable for a query, TofuBofu records that row as no_answer and drops it from the denominator rather than counting it as "did not name you". An engine going silent is not evidence you are absent, so it is excluded from scoring instead of dragging the Tofu Index down. If a query has too few real answers to score at all, its status reads "Too few answers to score" rather than a number.