AI visibility metrics measure how often and how well AI engines name your brand. The core set is visibility (how often you are named), share of voice (your share of all brand mentions), average position (where you appear), sentiment (how you are described) and citations (the sources behind the answer). This guide defines each one, shows how to calculate it and explains which numbers to trust most.
The numbers in the examples below are made up to show the arithmetic. They are not benchmarks.
What are AI visibility metrics?
They are the figures you get from repeatedly asking AI engines the questions your buyers ask, then counting what the answers say. Because answers can differ between runs, every metric is built from many checks, not one. If you are new to the idea, start with What is AEO? and the AI search glossary.
What is visibility (mention rate)?
Visibility is the share of checks in which your brand is named at all.
Visibility = checks that name you ÷ total checks
In an example, a brand is named in 18 of 60 checks, so its visibility is 30%. Always show the count beside the percentage, because "30% of 60" and "30% of 6,000" are not equally reliable.
A mention counts whether or not it includes a link. Visibility answers the most basic question: when buyers ask, are you in the answer?
What is share of voice?
Share of voice is your mentions as a share of all brand mentions across the same answers.
Share of voice = your mentions ÷ mentions of every tracked brand
In an example, across 60 checks the tracked brands are named 150 times in total, and yours is named 18 times. Your share of voice is 12%. Share of voice shows how you compare with competitors, not just whether you appear. A brand can be named often and still have a small share of voice if every answer also names four rivals.
What is average position?
Average position is where your brand appears in an answer when it appears at all. First is generally worth more than fourth, because buyers tend to read from the top.
Treat this one with care. SparkToro and Gumshoe.ai published research in January 2026 finding that the order of brands in repeated AI answers almost never matched, and the authors concluded that visibility measured across many runs is a reasonable metric while ranking position is not (SparkToro). The study is not peer reviewed, and one of its authors works for an AI tracking company, which the authors disclose. Still, the sensible reading is to use average position as a direction over many checks, and not to chase a single "rank 1" answer.
What is sentiment?
Sentiment is whether an answer describes your brand positively, neutrally or negatively. A mention that says "reliable but expensive" is not the same as one that says "the cheapest option." Sentiment matters most when visibility is already decent, because a negative description in the answer can cost you the buyer even when you are named. It also flags wrong or outdated claims worth correcting at the source.
What are citation metrics?
Citation metrics show which sources the engine used for its answer, and whether they include you. The useful ones are:
- Top cited domains, split into your own site, competitors' sites and third parties
- Citation gaps: sources that cite your rivals but never you
ChatGPT search shows the web sources behind an answer (OpenAI), and Google's AI features show supporting links (Google Search Central). Citations tell you why an answer went the way it did. See AI citation gaps.
Which AI visibility metric should you trust most?
| Metric | Answers | How much to trust it |
|---|---|---|
| Visibility | Am I named? | High, when built from many checks |
| Share of voice | How do I compare with rivals? | High, with a fixed competitor set |
| Average position | Am I named early? | Directional only |
| Sentiment | How am I described? | Good for spotting patterns and errors |
| Citations | Why did the answer go this way? | High as evidence, useful for action |
How many checks do you need?
The SparkToro authors suggest running prompts many times and averaging, in the range of 60 to 100 runs, and note that the right number for statistical confidence is still an open question. In practice that means a daily check on each engine, kept as history, so the total grows every week. Judge changes over weeks, and compare against your baseline. How to track your brand in AI answers shows how to set that up.
How should you slice the numbers?
Averages hide the useful detail. Break each metric down by:
- Engine: ChatGPT, Gemini, Google AI Overviews and Google AI Mode can differ.
- Subcategory: a group of related prompts, such as pricing or a product line. See How to choose the prompts to track.
- Wording: the same question phrased differently can change who is named. See Why AI answers change with wording.
- Time: week over week, against the baseline.
Where to see these metrics in Verte
Verte tracks visibility, share of voice, position, sentiment and citations on ChatGPT, Gemini, Google AI Overviews and Google AI Mode, with the count shown beside every percentage. See AI visibility tracking and citations and sources. Results arrive each morning from daily checks.
Start tracking your business or start as an agency. See current plans.