How the AI Shopping Visibility Index is measured

1. Buyer-intent query bank

For each category we maintain a bank of shopping questions built from the high-commercial-intent stages: Budget framingAttribute constrainedPurchase executionScenario confirmation Lower-intent stages (browsing, comparison, post-purchase) test something different and are handled separately.

2. Execution

Every query is sent to each AI engine. We capture whatever the engine returns — a product carousel, inline products, or a prose answer.

3. "Cited" — two surfaces

product = the store's product appears in a product card (ChatGPT, Copilot). mention = the brand is named in the answer text (Perplexity, Google AI Mode, which don't return shopping cards for these queries). Both count; both are labelled. Denominator = the number of query × engine checks.

4. Freshness

The bank is re-run monthly. Each page carries the month it was measured, and its figures are recomputed from that run.

5. Scope

Beauty › Skincare — ChatGPT, Copilot, Google AI Mode, Perplexity; 720 query-checks; May 2026; US. All 9 intent stages (Skincare dataset is smaller). 'product' = shown in a product card (ChatGPT, Copilot); 'mention' = named in the answer text (Perplexity, Google AI Mode).
Women's Fashion › Dresses — ChatGPT; 3778 query-checks; February 2026; US. 4 high-intent stages (Budget Framing, Attribute Constrained, Purchase Execution, Scenario Confirmation).

6. Which stores get a page

A store earns a page when the evidence is thick enough to say something true about it, or when an editor reviewed and approved it. Thick means at least 5 cited checks, cited by at least 2 engines, across at least 2 intent stages, appearing next to at least 3 other stores, and with a verified domain.

Data wrong for your store? There is a correction request on every store page.