Is your UGC cited by ChatGPT, Google & Perplexity?
Paste a product or collection URL. We grade the nine signals AI engines weight when they decide whose customer photos, reviews and UGC to quote: Product schema, named/attributable reviews, AggregateRating, AI-bot crawl access, llms.txt, agent-callable commerce tooling and more. The full result shows on screen instantly. Drop your email for a PDF report + a weekly re-run.
How we score
Nine signals, reduced from a 42-point internal scorecard.
The rubric started as the full 42-signal AEO scorecard we run on every Idukki onboarding. Two seasons of data narrowed it to the eight that explained 90% of the variance between a cited page and an uncited one. The other 34 still matter, they just do not move the result enough on their own to be worth a free-tool check. The ninth row is newer than the scorecard: agent-callable commerce tooling only became a live question in August 2026, when Shopify switched WebMCP on across every Liquid storefront, so it has no two-season history behind it yet.
- 01
Product JSON-LD present + complete
Without Product schema, AI engines fall back to scraping prose, which they cite far less often than structured data.
- 02
AggregateRating schema with count + value
The single strongest citation signal in our 42-signal AEO data.
- 03
Named, attributable Review block(s)
AI engines quote named reviews verbatim more often than they paraphrase prose.
- 04
robots.txt allows AI bot crawl
GPTBot, ClaudeBot, Claude-Web, PerplexityBot, Google-Extended, CCBot. If your robots.txt blocks any of these, those engines will not cite your page.
- 05
llms.txt present at site root
A signed table-of-contents for your catalogue. Cheapest single fix for AI visibility, and most stores do not have one.
- 06
Agent-callable commerce tooling (WebMCP)
Whether an AI agent can act on your store rather than only read it. Shopify has exposed WebMCP tools on every Liquid storefront since 21 August 2026, with no merchant setup. Tool-calling is not trust, so this row sits alongside your evidence signals, not instead of them.
- 07
Open Graph + Twitter Card tags
Agentic surfaces (Slack, WhatsApp, link unfurls) read these. Missing them = broken preview = lower citation rate.
- 08
Canonical URL declared
AI engines deduplicate variant URLs against this signal. Without it, variants compete with each other for citation share.
- 09
Meta description AEO-friendly
120-200 chars, fact-shaped, no marketing-tell openers. AI engines often quote the description verbatim in their answer snippet.
What this tool does not do.
We grade the signals AI engines weight, not the engines themselves. Actually querying ChatGPT, Claude, Perplexity and Gemini on a per-PDP basis costs roughly £0.40 per check; we keep the tool free by grading the upstream signals instead. Pass these nine and the citation rate follows. Fail them and no amount of waiting fixes it. The WebMCP row is the one exception to how the other eight read: we can confirm agent tooling where a platform ships it automatically or a page declares it inline, but WebMCP tools register at runtime in the browser and have no public registry, so a failing row there means not-detected rather than definitely-absent.
Pair this with the PDP audit for the full picture.
This tool grades AI citation readiness. The PDP audit grades UGC + conversion readiness. Together they cover the two questions every merchant asks first: will this page be seen, and will it convert.
FAQ
What does the AI visibility check actually test?
Nine signals AI assistants rely on when deciding whether to cite a page: Product schema, AggregateRating, named reviews, robots.txt access for AI crawlers, llms.txt, agent-callable commerce tooling, Open Graph tags, canonical URL and meta description. It grades the machine-readable surface of your page, which is what ChatGPT, Claude, Perplexity and Gemini can actually see.
Does it query ChatGPT or other AI engines directly?
No, and tools that claim to are selling snapshots: AI answers vary by phrasing, session and day. Grading the underlying signals is the durable measure, because pages with strong structured evidence are what the engines consistently cite.
My score is low. What should I fix first?
Crawler access first: if robots.txt blocks AI bots, nothing else matters. Then Product schema with AggregateRating, then named, dated reviews. The report orders your gaps by that priority.
How does the WebMCP row work, and how sure is it?
It answers a different question to the rest of the report. The other rows ask whether an engine can read and trust your page; this one asks whether an agent can act on your store without scraping it. Shopify announced WebMCP support for Liquid and Hydrogen storefronts on 5 August 2026, effective 21 August, and it is live on every Liquid storefront with no app install and no configuration, so confirming a Liquid storefront confirms the tools. Beyond that the honesty caveat matters: WebMCP tools are registered at runtime through a browser JavaScript API, with no well-known URL and no public registry, so a server-side fetch cannot enumerate them. On a non-Shopify stack we report not-detected, and you should confirm in a Chromium browser with the WebMCP origin trial enabled.
If Shopify gives me agent tool-calling for free, why do I need llms.txt?
Because they solve different halves of the problem. WebMCP lets an agent search your catalogue, edit a cart and start a checkout. None of that tells the agent your product is the one to recommend. The evidence an engine leans on when it picks between three plausible products is structured, attributable and specific: complete Product schema, real AggregateRating, named reviews, rights-cleared customer photos and video tied to the SKU. Shopify supplies the verbs. You still have to supply the reasons.
How is this different from the AEO brand report?
This tool grades one product page URL. The AEO brand report grades your brand's whole citable footprint: domain-level signals, review presence and machine-readable infrastructure.
Want to ship one of these?
Talk to us, we’ll get the right tool live on your stack.
Day-1 sample data on your live IG / TikTok / Reviews, plus a setup plan tailored to Shopify, WooCommerce, BigCommerce, Wix or any HTML site.
- No credit card
- Cancel anytime
- GDPR + SOC 2 (2026 roadmap)
