What AI assistants actually see on an online store
An audit that shows what AI readers receive from a shop, before and after fixes.
A working sample we built to show how we would solve a problem like this. It is not a delivered project.
In plain terms
AI shopping assistants read web pages differently from people. This tool checks what they actually receive from a store, and was run before and after fixing the store's page templates. On a test store built with typical problems, one content page showed only 12 of 64 entries to AI readers. Product details appeared twice, and the shop's bot protection blocked one assistant's reader entirely.
- Shows exactly what each AI reader receives.
- Finds hidden or duplicated information.
- Re-runs after the fixes to prove they worked.
What this does not claim: Run on a made-up test store with no real shop or client data. The tool only runs against sites you own or have permission to test.
Technical details
Audit tool with per-user-agent fetch tests. Test store reproduces a content hub exposing 12 of 64 entries in raw HTML, duplicate Product JSON-LD blocks, a hidden 'Sold out' label, and bot protection refusing AI crawler user agents. Three Liquid fixes rendered against fixture data, with a before/after comparison.
- Python
- Shopify Liquid (rendered with a Liquid engine)
- JSON-LD
Need something like this?
Tell us the problem in your own words. We reply with a plain, practical plan.
Discuss a project like this