Visibility Score is the primary metric used to quantify your brand’s presence in AI-generated answers. It is calculated as the percentage of your prompt set in which your brand appears at least once across a given AI system’s responses.

The calculation

Visibility Score = (Number of prompts that mention your brand / Total number of prompts run) x 100

If you run 50 prompts through ChatGPT and your brand appears in the response to 18 of them, your ChatGPT Visibility Score is 36%.

Most tools, including AskRank, compute this score separately per AI system so you can see your performance in ChatGPT, Claude, Perplexity, and Gemini independently. An aggregate score across all systems is useful for a headline number, but the per-model breakdown tells you where to focus.

What counts as a mention

A “mention” typically means your brand name (or a recognized variant) appears anywhere in the generated response. This includes:

  • Being named in a list (“tools like Notion, Linear, and [Your Brand]…”)
  • Being the subject of a recommendation (“I’d suggest checking out [Your Brand]…”)
  • Being mentioned as context or contrast (“unlike [Your Brand], Competitor X…”)
  • Being cited as a source or reference

Negative mentions count toward your Visibility Score because they indicate the model knows your brand. Sentiment is a separate dimension from mention rate.

Interpreting your score

There are no universal benchmarks for “good” Visibility Score because they vary by category. A niche B2B tool with 10 direct competitors might achieve a 60% score in its specific category prompts. A general-purpose tool in a crowded market might score 15% across a broad prompt set.

More useful than absolute benchmarks are:

  • Your own trend over time: is the score improving, stable, or declining?
  • Comparison across models: do you score 45% in Perplexity but only 12% in Gemini? That gap is actionable.
  • Prompt-level breakdown: which specific prompts are you mentioned in, and which are you missing?

Multi-sampling and score reliability

AI model outputs are not deterministic - the same prompt can produce different results on different runs. For this reason, serious tracking tools run each prompt multiple times and average the results. AskRank samples each prompt 1-5 times per run depending on your plan, which smooths out random variation and gives a more reliable score.

Score drops as alert triggers

One of the most practical uses of Visibility Score is as a threshold for alerts. If your score drops by more than 10 percentage points within a 24-hour window, something may have changed: a competitor announcement, a model update, or a shift in how the model is retrieving web sources. AskRank monitors for these drops and sends alerts via email or Telegram so you can investigate promptly.