Across the assistants shoppers use
We ask neutral buying questions, in many phrasings and with no brand names, across ChatGPT, Gemini and Perplexity, and record which brands each one names, and how often.
When a shopper asks ChatGPT, Gemini or Perplexity what to buy, a few brands get named and the rest don’t. We measure which brands they name, and how often, for real buying needs in your category. You see where you stand, how it moves from round to round, and, if you earn it, a badge to show on your product pages.
See the badge live on a demo store → a fictional Shopify shop
When a shopper asks an assistant what to buy, a handful of brands make the list and the rest don’t.
That list already works as a sales channel, but it is invisible and changes with every question.
What holds still enough to measure is not the ranking. It is how often a brand is named, and that is what we measure.
We ask neutral buying questions, in many phrasings and with no brand names, across ChatGPT, Gemini and Perplexity, and record which brands each one names, and how often.
Every answer is captured with its sources and a timestamp, and sealed. Each rate is tested against that need’s own chance rate, and the sources are checked for diversity. Every answer is sealed: the sealed tree and a sample of the answers are public, and the full set is available on request.
A brand that is named consistently, for a real buying need, earns a Verified record: on its page in our register and, through the Shopify app, on its product pages. Every surface links back to the evidence.
What lands in a store is a small, neutral badge that takes your theme’s styling and always carries its category, the assistants behind it, the date and a link to the full evidence.
We publish the method, not just the conclusion. Each record is measured across 3 web-grounded assistants and tested against a chance baseline, and every round measures it again. Every record links to its evidence: the hash of every observation, and a published sample of the answers.
| Category / intent | Engines | Null baseline | Top frequency | Trend | Status | Updated |
|---|---|---|---|---|---|---|
| hiking Xero Shoes | 3 / 3 | 32.5% | 100% | — | ◉ Verified | 11·03 |
| everyday minimalist Lems Shoes | 3 / 3 | 24.7% | 93% | — | ◉ Verified | 11·03 |
| wide feet | 3 / 3 | 31.4% | 94% | — | Measuring | 11·03 |
| running | 2 / 3 | 42.1% | 93% | — | Measured · no category yet | 11·03 |
| first pair | 2 / 3 | 31.1% | 85% | — | Measuring | 11·03 |
| office / dressy | 2 / 3 | 28.8% | 75% | — | Measuring | 11·03 |
Each need has its own chance rate. A record reads Verified only when a brand clears it with at least 3 of the 3 assistants, holds across the ways we phrase the question, and meets the rest of the published criteria. The worked example (Xero Shoes · hiking) reads Verified at 100% across 3/3 assistants, against 32.5% by chance. Every round checks it again, and a record that stops holding comes down. No brand appears as Verified before its evidence does.
Earned, not bought. See what AI recommends in your category. Start with a free check of your own.