See which Shopify products AI recommends — and fix the ones it skips.
Your merchandising flagged the best-sellers; ChatGPT is recommending three products you've never prioritized. eCommerce Insights tracks every variant across the six engines, scores each PDP twice, and lands the fixes as metafield and title diffs — with a push button for Shopify (Early access, Growth plan and up).
Your SEO stack watches the SERP. Your buyers moved to the answer.
Engines cite variants, not parents.
A men's merino base layer in medium is a different SKU than the large — different price, different availability, sometimes a different answer. ChatGPT Shopping and Perplexity cite them as distinct products. Tools that bucket variants into one parent lose the signal where the revenue is.
Theme JSON-LD ships with holes.
Most Shopify themes emit Product JSON-LD from theme.liquid or a snippet — usually without gtin, material, or a complete offers block, and apps can emit a second conflicting copy. Engines resolve the ambiguity by citing a competitor whose schema is complete.
Agentic checkout is being wired now.
Shopify's agentic storefront work and the checkout protocols (ACP from OpenAI/Stripe, UCP from Google) are in pilot as of mid-2026. The PDPs that win citations today are the same ones agents will be able to draft into carts — if the price, availability, and policies are machine-readable.
Two scores per product, in Shopify vocabulary.
The citation score explains why engines aren't recommending the product. The agent-readability score tells you whether a shopping agent could parse the PDP at all. On Shopify, both resolve to fields your team can edit this week.
| Check | Score | Shopify surface |
|---|---|---|
| Product JSON-LD completeness — offers, gtin, brand, material | Both | theme + apps |
| Variant resolution — distinct URL, price, availability per variant | Citation | variants |
| Structured attributes — GTIN, material, origin in renderable fields | Citation | metafields |
| Title and description answer coverage vs. buyer-phrased queries | Citation | product fields |
| Review signal — recency and Review/aggregateRating markup | Citation | reviews app |
| robots.txt admittance for GPTBot, PerplexityBot, ClaudeBot, Google-Extended | Agent | robots.txt.liquid |
| Machine-readable price, availability, returns and shipping policies | Agent | policies + schema |
| llms.txt presence and freshness | Agent | llms.txt |
Full methodology: citation score · agent-readability score · PDP Score docs.
Diffs that land as metafields, not tickets.
Every failing check produces a reviewable diff: a proposed title, description, bullet set, schema block, or metafield value, side by side with what's live. Recommendations write to metafields your theme already renders — GTIN, brand, material, country of origin — so there's no dev ticket and no Liquid rewrite. Your existing schema app keeps shipping its JSON-LD; eCommerce Insights fills the holes it couldn't.
A merchandiser reviews, edits, approves. On the Growth plan and up, the approved change pushes to Shopify through the Admin GraphQL API — logged, reversible, attributed to the approver. Shopify push · Early access Everyone else exports a CSV their team can ship today. Details on the diff workflow: PDP optimization.
One product, before and after.
Illustrative, not a customer claim: a 200gsm merino base layer whose theme JSON-LD lacked offers.availability and whose title didn't carry weight, gender, or use case. The diff added both metafields, completed the offers block, and rewrote the title in the buyer's phrasing. The next two scan cycles recorded the SKU entering ChatGPT Shopping cards and the Perplexity citation set for "best merino base layer for cold-weather running."
Numbers above are illustrative. eCommerce Insights is in early access and publishes no case-study numbers it cannot stand behind.
Related pages.
Product-level tracking
The tracking primitive everything else builds on. Every variant, every engine, every refresh.
Guidellms.txt for Shopify
Serving an llms.txt from a Shopify store without forking the theme.
ChannelFor headless
Running Hydrogen or a custom storefront? The rendered-audit workflow.
Also: the product AI visibility guide · best AI visibility tools for Shopify · how SKU-level differs from Profound.
Frequently asked questions
Does eCommerce Insights require Shopify Plus?
No. It works on every Shopify plan from Basic upward. Plus merchants get higher SKU limits on the Growth and Agency & Enterprise tiers plus SSO and a DPA on request. The per-SKU tracking, two scores, and PDP recommendations are identical on every Shopify plan.
Can it push changes to my store, or do I export?
Both. On the Growth plan and up, approved diffs push to Shopify through the admin API (Early access): title, description, and metafield writes, each logged, reversible, and attributed to the approver. Every plan can export recommendations as CSV and ship them through your own workflow.
How does it handle variants versus parent products?
Variant-level by default. A medium and a large of the same base layer are two SKUs, because ChatGPT Shopping and Perplexity cite them as distinct products with distinct prices and availability. Rollup views show the parent-product picture when you triage at the collection level.
Can I keep my existing schema app, like Yoast or Schema App?
Yes. eCommerce Insights audits the JSON-LD your storefront actually serves, whatever emits it — the theme, an app, or hardcoded Liquid. Recommendations are incremental: fill the missing Product fields, add the missing offers block. Nothing asks you to replace the app you already trust.
What data does it read from my Shopify admin?
Product titles, descriptions, variants, images, collections, tags, and metafields in the namespaces you grant. No customer PII, no orders, no checkout data. Scopes are read-only unless you explicitly enable the write scope that lets approved diffs push back to the catalog.
Does it work with headless Shopify or Hydrogen?
Yes. Catalog reads come from the admin API, so a Hydrogen or custom storefront doesn't change what gets audited. Scoring runs against the public PDP that AI engines actually crawl — rendered with JavaScript when needed. The headless solutions page covers the rendering workflow in detail.
Ask AI about eCommerce Insights for Shopify
Have your preferred AI engine summarize this page for your store.
Shopify runs the store
eCommerce Insights runs the AI visibility.
Connect from Shopify admin in five minutes. First scan within 24 hours.
Starter $99/mo · Growth $349/mo · 14-day trial