How we check, and what we can't know
We want you to trust the report because you can see how it was made, not because of big claims.
What we read
Only public pages: your Shopify catalog (/products.json, up to 250 products), collections, homepage, About page, shipping and returns pages, robots.txt and three sample product pages. We never log in, we crawl politely and our own reader follows your robots.txt.
How we write shopper questions
We pull your category, audience, main promise, price band and your own branding words from the store. Then we rewrite them the way a buyer who has never heard of you would ask, for example “best self-care gift box under $50”. Questions never contain your store name or domain, because a branded question only proves the AI can search for a name. Every question is listed in your report.
How we ask the assistants
We use the official APIs of ChatGPT (OpenAI), Gemini (Google), Claude (Anthropic) and Perplexity, with web search turned on. In the Audit each question is run twice on each assistant, and we show every answer word for word with the model and the time.
The findability score
Out of 100, from checks we can verify: AI crawlers allowed in robots.txt (20), Product schema (20), price and availability fields (15), brand field (10), GTIN or MPN (10), descriptive product copy (10), shipping and returns pages (10), and a readable About page (5). It measures how readable your store is to AI, not how often you are recommended.
Honest limitations
- No search volumes. None of the AI companies publish how often people ask a question or how often a store is recommended. Nobody has this data, so we don't show it.
- Answers change between runs. The same question can get a different answer an hour later. That is why the Audit runs each question twice and Monitoring tracks the trend over weeks.
- API answers are close to, not identical to, the apps. The consumer ChatGPT app also uses Shopify Catalog shopping results, personal memory and location. We don't automate the consumer chat apps.
- Our questions are realistic, not real. They are written from your store, not taken from real shoppers' chats. In the Audit you can add 2 of your own.
- Your positioning is our reading of your pages. If it looks wrong, that itself is a signal: AI reads the same pages.
- Fixes improve your odds, they don't guarantee placement. No one can promise an AI will recommend you.
Independence
AsAISeesIt is independent and not affiliated with OpenAI, Google, Anthropic, Perplexity or Shopify.