Understanding your AI visibility
How to read AI Visibility's prompt-by-prompt citation results and act on gaps.
Updated July 19, 2026
AI Visibility's results live in a prompt-by-prompt citation table, not a single blended score. This is how to read it.
Where to see your results
Go to AI Visibility under Analysis. You'll land on Prompt Results, showing how many of your tracked prompts have been evaluated so far, for example, 10/10.
How a result is structured
Each row is one prompt, the exact question a buyer might ask an AI. Two columns summarize it at a glance:
- Best Tier, your strongest result for that prompt across every engine
- Category, the topic cluster it belongs to
Expand a row and you'll see four independent results, one per engine: Perplexity, ChatGPT, Gemini, Claude. Each is marked Cited or Not cited with its own tier. A prompt isn't simply won or lost; it can be cited on Perplexity and invisible on Gemini in the same breath, and that split is the whole point of looking here instead of at one number.
Filtering the table
Six tabs sit above the table: All, Cited, Gaps, Category, Brand, Discovery. All and Cited are self-explanatory. The other four are for narrowing down fast, by prompts with no citations anywhere (Gaps), by topic (Category), by which competitor is winning a given prompt (Brand), or by what's new since your last check (Discovery).
Reading the domains cited
When an engine cites something, its full ranked source list appears underneath, #1 through however many domains it pulled from for that prompt. This is where "citation opportunity" and "content gap" stop being abstractions: you're looking at the exact sites a specific engine trusted for a specific question.
Why content impact needs a connected domain
A banner on this page reads: "Connect a publishing domain to measure content impact." Until your custom domain is verified, AI Visibility can still show you who else is cited, but it can't tell whether content you've published through it is showing up in these results. Domain verification isn't only a publishing requirement, it's the only way to close the loop on whether publishing is working.
Acting on a gap
A prompt with no citations anywhere, or a domain list that never includes you, is your clearest opportunity. Start with gaps in your highest-value categories and use the Overview dashboard's "Generate content" action directly against them, rather than treating every gap as equally urgent.