How the AI Shopping Visibility Index is measured
Every figure in the index comes from questions we asked AI shopping assistants and the answers they gave.
1. A bank of buyer-intent questions
For each category we keep a bank of shopping questions, each tagged with the buying stage it represents. Each question is counted once per engine, even if the same wording appears in the bank twice.
2. Asking the engines
Every question is sent to each engine the category is measured on. We record what the engine returns: a product card or a store named in the text of the answer.
3. What counts as cited
- ChatGPT: product card
- Copilot: product card
- Google AI Mode: text mention
- Perplexity: text mention
Visibility is the share of query-checks (one question asked of one engine) where the store was cited. Average position is the mean slot the store held in the answers that named it with a position; #1 is first.
4. Which stores get a page
A store gets its own page when it is cited in at least 5 checks in a category's latest run. A brand with several domains gets one page, on the domain cited most. Every other cited store is listed on the leaderboard.
5. Freshness
Each category is measured once a month. Every page shows the month its figures come from.
6. What this run covers
- Beauty › Skincare: 180 questions × 4 engines = 720 query-checks; ChatGPT, Copilot, Google AI Mode, Perplexity; May 2026; US. Stages: Problem recognition, Category exploration, Comparison, Budget framing, Attribute constrained, Purchase execution, Scenario confirmation, Validation / risk reduction, Post purchase.
- Women's Fashion › Dresses: 3778 questions × 1 engine = 3778 query-checks; ChatGPT; February 2026; US. Stages: Budget framing, Attribute constrained, Purchase execution, Scenario confirmation.
Data wrong for your store? Email [email protected] or use the correction link on the store's page.