AI citation index
ReserveBeat's AI citation index measures which web domains AI answer engines cite when asked neutral stablecoin reserve questions. Each week the same versioned, brand-free buyer queries run against a pinned set of engines, and every cited domain is counted from stored responses. Figures are observation counts — never rankings, endorsements, or reserve figures.
By Aaron Cohen
No citation runs have landed yet. The index appears here after the first configured weekly run stores its observations — nothing is estimated or backfilled in the meantime.
Frequently asked questions
What does the AI citation index measure?
It counts which web domains AI answer engines cite when asked a fixed set of brand-free questions about stablecoin reserves. Each weekly run stores every response verbatim, and the counts shown here are computed from those stored rows — observations of engine behavior, not quality judgments about any site.
Why are the queries brand-free?
A query that names ReserveBeat cannot measure unaided visibility — an engine told the brand name will tend to repeat it. The pinned queries ask neutral buyer questions, so a brand mention or citation only counts when an engine surfaced it on its own.
Is the weekly summary written by a person?
No — it is machine-generated by a pinned model from the run's stored rows, and it is disclosed as such. It passes the same forbidden-term copy checks as every human-written string before it is stored; if it fails, no summary is stored at all.
Are these reserve or coverage figures?
No. Nothing on this page is a reserve figure, a coverage ratio, or a statement about any issuer. The index observes how AI engines answer public questions — it is a measurement of the information landscape, not of reserves.