"Do I appear in AI?" can't be answered by gut feeling — nor with a single question to a single model. The answer only becomes reliable when you measure systematically and reproducibly. That's exactly what Surfaced uses the 17×4 method for.
17 questions × 4 engines = 68 answers
The starting point is 17 real customer questions — the phrasings potential customers actually use to search for the service, tailored to industry and region. Each of these questions is asked across four AI engines: ChatGPT, Perplexity, Google AI and Claude. That yields 68 individual answers per language.
From the answer set to the score
Each of the 68 answers is evaluated by two criteria: is the company named? And is what the AI says about it correct? The mention only counts if both are true. The visibility score is the share of these valid mentions across all 68 answers, weighted by each engine's reach.
The key is repeatability: the same measurement runs identically every month and is evaluated manually rather than by tool automation — so a snapshot becomes a comparable trend from which real development can be read. In multilingual markets, the full run happens separately per language.
You can apply the method yourself with a simple spreadsheet. Whether the AIs then actually send you visitors is shown by Surfaced's free plan — measured on your own website.
Test the dashboard free — from €0