AI assistants can reach your store. Can they read your products?
Shopify opened native category attributes so ChatGPT, Copilot and Meta AI can read your products. In most stores, those fields are empty. We fill them — with your own verified product information, never invented.
Powered by the Semantic Commerce Engine — part of the Semantic Commerce Layer™
How the Optimizer turns human-written product data into structured intelligence
The Optimizer works with the data it receives. Enriched product data gives it more verified context — and produces a stronger result.
Click to enlarge Real screens from an actual merchant in Shopify.
The drawers are open. Most of them are empty.
Shopify opened its merchant catalogs to AI assistants, and millions of stores became reachable overnight. The shelf space is real — but AI agents don't read marketing copy. They read structured attributes: category, material, fit, capacity, compatibility, use case.
Most catalogs were written for humans browsing a page, not for machines choosing a product. So the assistant can't tell what the product is, who it's for, or when to recommend it — and it recommends something else.
That gap isn't a design problem or an ads problem. It's a data problem, and it's fixable.
Built for Shopify stores that already sell well to humans.
- Shopify merchants whose category metafields are empty or partially filled
- Catalogs up to 1,500 SKUs (larger catalogs: talk to us)
- Stores where product content lives in tabs, accordions or apps — invisible to machines
- Agencies preparing client catalogs for agentic commerce
Not a fit if: your store isn't on Shopify (ask us about AI-Ready Commerce Feeds™ instead), or you need copywriting for shoppers — this is structured data for machines.
What we need from you.
- 1.Your Shopify store connected through a custom app, with the minimum permissions required to read your products and write the fields we manage. Nothing else.
- 2.Your existing product content — titles, descriptions, specs, variants, materials, warnings. Whatever you already published is the raw material.
- 3.Your approval before anything is written to your store.
That's it. No spreadsheets, no re-platforming, no theme changes.
What the Optimizer does.
Resolves the right Shopify category for each product — against the official taxonomy, not a guess. When the evidence isn't conclusive, it doesn't invent: the product goes to human review.
Fills the category attributes Shopify exposes to AI assistants, using values the platform accepts for that category.
Builds a machine-readable description of each product from your published evidence — what it is, what it's for, who it's for, what it's made of — plus SEO title and meta description aligned with the same verified facts.
Leaves a report per product: what was written, what was skipped, and why.
Rule we don't break: Verified, never invented. If a fact has no evidence in your catalog, the field stays empty. Warnings and compliance text are preserved verbatim.
What you get.
- Official Shopify category assigned per product
- Category metafields populated with accepted values
- A machine-readable product description written from your evidence
- SEO title + meta description per product
- A per-product report: written / skipped / reason
- A re-measured Agentic Catalog Readiness Score™ — before and after
Everything lands inside your store. It's your data, on your platform. If you ever stop working with us, nothing is taken back.
How we handle access to your store.
We ask for the minimum permissions needed to read your products and write the fields we manage — and nothing else. We never request access to orders, customers or checkout.
- Least privilege: only the scopes this product needs
- Credentials encrypted at rest, never shown in logs, exports or support tickets
- Every write is logged: which store, which products, which fields, when
- Revocable anytime: uninstall the app and access ends immediately
- Extended permissions (like writing new catalog values) require your separate written approval
Your catalog is your business. We treat access to it accordingly.
Three steps.
- Step 1 — Measure
Run the free Agentic Catalog Readiness Audit™. You see the gap before spending anything.
- Step 2 — Remediate
Initial Remediation runs across your catalog. If the source content is too thin to work with, Optimizer Human Assistant steps in first: a human confirms category, title and description, then the Optimizer runs again on better material.
- Step 3 — Measure again
The same audit runs on the same catalog. The change is measurable, product by product.
Optional, after that: Care keeps watch — new products, edits, taxonomy changes, drift.
Measure → remediate → measure again.
- Before
- category unassigned · 0 category attributes filled · description written for humans only
- After
- official category assigned · attributes filled from verified evidence · machine-readable description · readiness score re-measured
What we claim: your catalog becomes readable and interpretable by AI systems, and we can show the difference with the same measurement before and after.
What we don't claim: we don't promise rankings, recommendations or sales. Those are different outcomes and they need different measurement.
What Shopify merchants ask.
- Where do the attributes come from? What stops the AI from making things up?
- Everything comes from your own published product information — titles, descriptions, specs, variants, materials, warnings. That evidence is what we work from. If a fact isn't supported by your catalog, the field stays empty rather than being filled with a guess, and warnings are preserved word for word. Verified, never invented.
- Which Shopify fields do you modify, and which do you leave alone?
- We write only the surfaces this product manages: the official Shopify category, its category metafields, a machine-readable description field, and SEO title and meta description. We do not touch your visible product titles, your product body, prices, inventory, promotions, tags or URLs.
- How do you decide the right category? What happens with ambiguous products?
- Products are resolved against Shopify's official taxonomy using the evidence in your catalog. When that evidence isn't enough to resolve a category confidently, we don't guess — the product goes to human review before anything is written.
- What if Shopify rejects one of the fields? Does the whole product fail?
- No. Writes are isolated from each other: the category is set first, then each attribute is written independently. One rejected attribute doesn't undo the rest of the product or the batch, and the report tells you exactly which fields were accepted, which weren't, and the reason returned.
- How long does this take?
- Days, not weeks — and you approve before anything is written to your store. Initial Remediation runs across the catalog in one pass; what sets the pace is the review, not the processing. If your source content needs repair first, Optimizer Human Assistant adds a step before that pass.
- My product content is thin. Is this going to work?
- It works with what's there, and it tells you the truth about the rest. Thin source content produces an honest minimum, not invented filler. When the source itself needs repair, Optimizer Human Assistant fixes the source first — a human confirms category, title and richer description — and then the Optimizer runs again on better material.
- Do I review anything before it goes live?
- Yes. Nothing is written to your store without your approval, and you keep a per-product record of what was written and what was skipped. Your own edits stay yours: a value your team corrects is treated as authoritative, not as something to overwrite.
- Why should I give you write access to my catalog?
- Because the work happens inside your store, and you keep control of it. We operate under least privilege: the minimum permissions to read your products and write the fields we manage, never orders, customers or checkout. Credentials are encrypted and never appear in logs or exports, every write is recorded, and you can revoke access at any time by uninstalling the app. Anything beyond the standard scope requires your separate written approval.
- What happens after the remediation is done?
- Your catalog stays as remediated — the work lives in your store and it's yours. Catalogs drift, though: you add products, edit descriptions, and Shopify updates its own taxonomy. Care is the optional monthly layer that watches for exactly that and keeps the catalog from sliding back.
- Do I need a specific Shopify plan for this?
- No. Category attributes are part of Shopify's product data on every plan — they're the same fields whether you're on Basic or Plus, and on most stores they're sitting empty regardless of what you pay Shopify. What changes with your plan is other things; this isn't one of them.
See what your catalog looks like to a machine.
The Agentic Catalog Readiness Audit™ is free. Start there — you'll see the gap before you spend anything.
Where this sits
- Part ofthe Semantic Commerce Layer™NSOLVIA's framework for practising IEO on commerce data — the interpretability layer between catalogs and the systems that read them.
- Goes deeper
- how the Optimizer writes verified data into Shopify — What the Optimizer writes, what it never touches, and where the attributes come from.
- Alongside it
- the free catalog readiness audit — The free instrument: reads a real product the way machine systems do and reports its interpretability baseline.
- AI-Ready Commerce Feeds™ — Enrich the catalog once and deliver it in the format each destination expects.
- Catalog Export™ — The enriched catalog delivered as a structured dataset you own, with no subscription attached.
NSOLVIA Agentic Catalog Optimizer™ for Shopifysits inside NSOLVIA's interpretability stack — an open discipline, the framework that implements it, and the instruments that apply it.