AI visibility has three parts: whether an answer names you at all, how it describes you (as a top pick, in passing, or with a caution), and whether it cites your website or someone else’s when it does.
Because AI answers change from one ask to the next, visibility can’t be read from a single answer. It has to be measured over many answers to a fixed set of questions. At Citationly we report it as a count with its population named: “named in [n] of [N] answers we recorded”, for a stated set of questions, a stated model and date, with web search on. We never turn it into a score out of a hundred, because a score hides how many answers it rests on.
AI visibility is also always relative to the questions asked. A brand can be named in nearly every answer about one product line and in none about another. That is why a useful report breaks results down by product line and buying stage rather than giving one number.
Read how to measure AI visibility without a made-up score and why ChatGPT gives a different answer every time.