Skip to content

How to measure AI visibility

How to measure a brand's visibility in ChatGPT, Gemini, Claude and other AI assistants reliably: questions, repetition, engines and metrics.

Last updated:

Short answer

AI visibility is measured by asking the questions your customers would ask to several AI engines, several times, and recording for each answer whether your brand appears, in what position and tone, and which sources are cited. A reliable measurement repeats, reports each engine separately, never counts an unmeasurable case as zero, and states how uncertain the result is.

Four core metrics

  • Mention rate: in what share of questions the brand appears.
  • Position: where it ranks among the recommended brands.
  • Tone: positive, neutral or negative.
  • Citation share: how much of the answers' source lists are your and your competitors' sites.

Rules of a good measurement

  • Write questions the way customers do; keep the brand name out of recommendation questions.
  • Ask every question again: answers change from run to run.
  • Measure engines separately: ChatGPT, Gemini, Claude, DeepSeek and Grok look at different sources.
  • Keep answers with and without web search apart; they measure different things.
  • Do not count a missing answer or an error as "not mentioned"; that is missing data, not zero.
  • Give the result with a confidence range: 30% of 20 questions is not as certain as 30% of 200.

From measurement to action

The value of a measurement is in what it tells you to do. Knowing which sites are cited in answers that mention a competitor but not you (the source gap), and what the model searched before answering (background searches), directly steers content and outreach work.

Frequently asked

How often should I measure?

Weekly or every two weeks is usually enough to see change; measure again after a major content push.

How does Mavzena measure?

The full method is on the Methodology page: question generation, engines, reading answers, the score and its limits.

Sources

  1. A Critical Survey of Generative Engine Optimization (2023-2026)
  2. Aggarwal et al., GEO: Generative Engine Optimization (2023)

Does your brand show up in AI?

See what ChatGPT, Gemini, Claude and other engines say about your brand with a free audit.

No account needed. We read your page the way AI crawlers do, and you see what they can and cannot read.