AI rank tracking

AI rank tracking tools run the same prompt on a schedule against surfaces like ChatGPT, Perplexity, Gemini, and Google's AI Overviews and AI Mode, under fixed conditions — one locale, logged out, one device — then log four things: whether a brand appears at all, whether it is named, whether a URL is cited as a source, and what share of the generated answer it occupies against named competitors. Repeated over weeks, that produces a trend line the way a keyword position does, but the position itself is a proxy, not a slot. What it is not: AI visibility monitoring, the neighbouring term in this glossary, only asks the binary — was the brand named or cited at all. Rank tracking asks for standing within that citation: share of voice, inclusion frequency across a prompt set, or which specific URL got cited. Where it breaks: a generative answer has no ordered list, so there is no position 1 through 10 to hold. Answer share and inclusion rate move with locale, personalization and the underlying model version, so a single query is a sample, not a rank — tools that publish a stable number are averaging repeated runs, not reading one.