AI visibility
Brand citations on ChatGPT: read them, don't trust them
How to use RankAI's FAQ and aiVisibility page to read brand citations on ChatGPT and other AI engines without mistaking them for a guaranteed position.
If you're wondering how to read brand citations on ChatGPT without mistaking them for a promise of position, here's the practical answer: on RankAI's aiVisibility page you observe whether and how the brand appears in the answers of the named AI engines, while the FAQ tells you what that data means and what it doesn't. Picture opening aiVisibility after a week of questions about your sector: you see the brand cited in some answers, missing in others, and in one case the measurement shows as N/A. Read it as a description, not a prediction: you're looking at an observation, not a guaranteed ranking.
FAQ and aiVisibility: two separate uses
The FAQ answers the question "how do I interpret this number?" The aiVisibility page answers the question "what was observed?" Keeping them apart avoids the most common mistake: treating a citation as if it were a position you already hold. RankAI shows observations on the AI engines named on the public pages: ChatGPT, Perplexity, Gemini, Grok and Google AI. Claude and Grok are measured only after you connect them. Gemini and Google AI Overviews are not the same observation, so don't add them together or compare them as if they were a single figure.
Why N/A is not zero
When a visibility figure is unavailable, RankAI shows it as N/A, never as zero. That changes how you read it: N/A means the data isn't available in that observation, not that the brand is invisible. Confusing N/A with zero leads to wrong conclusions about brand citations on ChatGPT and the other engines observed.
An operating method proposed by RankAI
What follows is an operating proposal from RankAI, not a universal standard: it's meant to help you read citations without turning them into expectations about ranking.
- Start from the question, not the number. Checkable rule: for every citation you notice, identify the question or context that produced it. If you can't reconstruct it, that citation isn't readable as a signal yet.
- Keep engines separate. Checkable rule: for each observation, note which engine is involved, remembering that Gemini and Google AI Overviews are not the same observation and that Claude and Grok appear only after you connect them.
- Treat N/A as missing data. Checkable rule: every time you see N/A, write "not available," not "absent." If your note says zero, correct it.
- Use the FAQ for meaning, not for prediction. Checkable rule: after reading the FAQ, write a sentence describing what you observed, without verbs like "will be" or "guarantees."
What this piece adds is a reading discipline: a citation is an event observed in an answer, not a position title. In practice, keep a log where each row holds the date of the observation, the engine involved, the starting question and the outcome (cited, not cited, N/A). This format has a concrete advantage: it makes missing data visible instead of letting it disappear into an average. If one week shows many citations and the next shows N/A on an engine, the log stops you from reading the N/A as a collapse. Separating rows by engine also keeps you from attributing to ChatGPT an observation that belongs to Perplexity or Google AI. Consult the FAQ before writing conclusions: it reminds you that Gemini and Google AI Overviews are not the same observation and that Claude and Grok only come in after you connect them. Finally, read the log as a snapshot of the past you observed, not as a forecast: no row lets you say the brand "will be" cited tomorrow. It's a working method, not a guarantee.
A limit to keep in mind
Here's the checkable limit: the citations you read in aiVisibility describe only the observations available on the named engines, and where the measurement is not available the product shows N/A. None of these observations is a guarantee of position, present or future. For the full picture on how to read this data, see the in-depth page.