CitedOS computes the standardized AI-visibility metric set for every prompt × engine, over multiple runs. Here is what each one means and how to read it.
The five metrics
Are you mentioned at all?
The percentage of runs in which your brand appears anywhere in the answer. The first question: do the engines know you exist for this prompt? Read low presence as a discoverability problem.
How much of the conversation is you?
Your mentions divided by all brand mentions across the answer set. SoV is competitive — it falls when rivals are named more, even if your own presence holds. The headline number for "who owns this category in AI".
Where in the answer?
Your rank or order when you are mentioned. Being named first reads very differently from being the seventh option listed. Lower (earlier) is better.
How are you framed? (0–100)
How positively the answer frames your brand, scored 0–100. A high presence rate with low sentiment means the engines name you — but not flatteringly.
How often, per N runs?
Mentions per N runs — the frequency view that complements presence. Useful for spotting prompts where you appear sometimes but unreliably.
How to read them together
- Low presence + low SoV → a discoverability gap: you're not in the retrieval pool. Start with the source gap.
- Good presence + low SoV → rivals dominate the same answers. Compare head-to-head in the dashboard.
- Good presence + poor position or sentiment → you're cited but framed weakly — a selectability and content-quality problem.