Methodology
This is exactly how Emphanic measures AI visibility. No black box, no prettier story than the data allows.
What we do
We ask 12 localized customer questions to four AI engines: ChatGPT, Claude, Perplexity and Gemini, all with web search on. The questions are phrased the way real customers ask them, for example 'what is a good physiotherapist in Breda?' or 'top 3 bakeries in Utrecht?'.
In every answer we check whether your business is mentioned, including spelling variants of your name. We also extract which other businesses get recommended.
How the score is calculated
Per engine we calculate the share of answers that mention you, from 0 to 100. Your total score is the average across engines. Mentioned in 6 of 12 Gemini answers means a Gemini score of 50.
Questions an engine failed to answer are excluded, so a technical failure does not distort your score.
Why results vary
AI models are not deterministic: the same question can produce a different answer tomorrow. The web sources engines rely on change constantly too.
A single measurement is therefore a snapshot. The honest measure is the trend: repeated measurements over weeks. That is exactly what the paid subscriptions will do weekly.
What we do not measure
We do not measure Google rankings, sentiment of mentions, or revenue. We count mentions in AI answers to realistic customer questions. Nothing more, nothing less.
We report what we measure, even when it is disappointing. Anyone promising you a guaranteed position in ChatGPT is not being honest with you.