Our Approach to Gathering AI Visibility Data
We process millions of real prompts across AI Mode, ChatGPT, Copilot and Perplexity, then track which brands, products, and domains show up in the answers. That's the raw data behind every score in the Index.
Real Intent Capture
This data results from processing millions of real user prompts. We use advanced semantic and agentic processing to categorize them and define intent, filtering out noise to focus on high-value queries.
Global Model Interrogation
We run these prompts across the tracked models and across every market we cover, so each market is scored on its own answers rather than on a global average applied everywhere.
Attribution Mapping
We analyze the generated answers to capture the "Kingmaker" domains: the specific sources models cite most often in the answers that name a brand, so you can see which sources the models lean on where you appear and where you do not.
