Methodology
How we measure your AI visibility — with the method in plain sight.
AI never answers the same twice. Any serious measurement has to account for that — here is how we do it.
What we measure
Your visibility percentage: in how many of ~25 real market questions your business appears when we ask the engines. Questions are written the way a real owner or customer asks — native-language, not translations — and stay fixed month to month, because changing them would break the comparison.
How we read
- Consumer fidelity. We ask the way your customer asks: clean session, no tricks, no specialist prompts.
- 3 passes per question, per engine. Because AI never answers the same twice. Three reads turn a blurry snapshot into an honest range.
- Result as a range, never a point. We report “31–36%,” not “34%.” And we only claim movement beyond the previous range — anything else is noise.
visibility % (range across passes)
3 reads per question — AI never answers the same twice.
Methodology visible in every report: your market’s real questions, native-language, movement claimed only beyond the range.
Which engines we read
Free read: Google only — its AI Overviews and its organic search. We say it plainly: the free analysis does not read ChatGPT.
Paid reads (Diagnostic and Monthly): the four main engines, plus Google.
3 reads per question, on every engine.
What we don’t do
Entiende. Arregla. Analiza.
Understand
- Your human communication and your technical infrastructure — the two readers.
Fix
- What blocks you, prioritized; pages that answer your market in your voice.
Analyze
- The full read again — same set, same engines, 3 passes — and the honest delta.
This is how the visibility line works → AI Search Positioning