Skip to main content
AI Visibility summary and platform metrics in the current product
Open Analytics → AI search → AI Visibility to compare how your brand appears in tracked answers. The screenshots use illustrative local data. Keep the prompt, group, category, and date scope consistent when comparing results.

Read the summary

Visibility, LLM position, and Citation share measure different outcomes. Visibility is a position-weighted score, not mention percentage. Average position includes a penalty for missing observations. Citation share measures your brand’s share of the sources attached to answers. See AI search metrics for definitions and examples. These metrics describe tracked answers; they do not measure website visits or predict conversion changes by themselves.

AI Visibility tab

Use the platform chart and table to compare where the brand has stronger ranked exposure. Switch chart views to inspect the trend or compare platform values. A weak platform should lead to a review of its saved answers, not an assumption about the cause.
Platform metrics with AI Visibility selected

LLM position tab

Compare average rank across platforms. Lower is better, but check visibility alongside it: missing observations contribute to the comparison metric. An unavailable position should not be read as a top rank or a measured zero.
Platform metrics with LLM position selected

Citation share tab

Compare the share of each platform’s citations that point to your brand. Open the Citations report to inspect individual pages and domains. A citation is a source reference, not a referral click.
Platform metrics with Citation share selected

Recent chats: All

Review the recent run list, including processing states. Check the prompt, engine, country, time, and top results. Waiting or unfinished answers are not evidence of missing brand coverage.
Recent chats with All selected

Recent chats: Mentioned

Focus on completed answers with a positive ranked position for your brand. Open useful examples to understand the language, placement, and evidence supporting the result.
Recent chats filtered to Mentioned answers

Recent chats: Missing

Focus on completed answers where the brand lacks a positive ranked position. Check the full answer and saved top results before choosing a content update. An incidental name mention can exist without ranked inclusion.
Recent chats filtered to Missing answers

Open an answer

Select a completed run to inspect its saved answer, ranked results, and citations. Use the saved run’s evidence when comparing outcomes; rerunning a prompt can produce a different answer. Pending runs may not yet have an answer to inspect.
Saved AI answer with run navigation, citations, captured queries, and other search sources
The evidence panel distinguishes Citations from Query fan-out and Other search sources. Sources labeled Not cited were returned but are not marked as cited in the answer. Do not add them to citation counts. Search details depend on captured data and subscription access; an absent section or upgrade prompt is not proof that the model did not search. Query cards open a new Google search, not the original search results. To compare historical evidence, use the saved runs. Opening a query-table row starts with its newest matching run; when several runs match, navigation above the answer lets you inspect the others without rerunning the prompt. See Query fan-out for the map, table, and run-level source limitations.

Choose the next action

  • Review a repeated missing pattern across relevant prompts before adding work.
  • Use the prompt and group filters to isolate a business-relevant segment.
  • Compare a changed platform with its citations and competitor results.
  • Check referral traffic separately in AI search Overview.

Where to go next