Methodology · Reviewed 2026-09-27

How Uprity measures AI visibility

The score is a directional sample of how configured AI assistants answered a controlled set of local-discovery prompts. It is not a search-engine ranking, market-share measurement, or guarantee that another user will receive the same answer.

Prompt sample

The free preview asks one discovery question. A paid scan asks up to three category-and-location discovery questions and one branded question. The business name appears in the branded question, so it does not affect the score.

One configured provider

A scan uses one available provider configured for web search. The dashboard shows its name, model, answer coverage and available citation-page URLs. Requested web search does not guarantee that every answer cites a page.

Score calculation

The 0–100 score is the percentage of successful discovery answers in which normalized name matching detects the business. Failed answers, branded questions, answer position, sentiment and citations do not add points.

Comparability over time

The trend shows scans using this score method. Compare similar prompts and the same provider and model; model updates, web results and answer randomness can still change the sample.

Important limitations

  • Mention detection uses normalized name matching and can miss aliases or confuse similarly named businesses.
  • Position is an estimate from list structure or answer location; it is not supplied by the provider.
  • Available citation links show pages referenced by the provider, not proof that the pages are accurate or mention the business.
  • Generated answers can be incomplete, outdated or wrong. Verify material decisions independently.

Use the score as a trend, not a promise

Review the underlying prompts, snippets, citations, and configuration alongside the headline number.