Your store is invisible to AI.
AI doesn't read your store the way a human does. A beautiful product description is useless to an agent: it needs the facts in structured, machine-readable form. Without them your product never makes the shortlist, even when it is the best option.
Waterproof outdoor jacket for everyday and trail.
{
"price": 189.00
"availability": "InStock"
"gtin13": puuttuu
"material": puuttuu
"returnPolicy": puuttuu
}Three routes to the AI
The market isn't ready, and that is your advantage. The routes don't exclude each other; they are built from the same data.
Shopify's own pipeline: store → Shopify Catalog → ChatGPT, Gemini and Copilot. Metafield data is fed to agents over the ACP protocol. US and CA stores for now; the EU follows as regulation moves.
Perplexity, Claude, Gemini and ChatGPT read structured data straight from your pages. Google Rich Snippets and E-E-A-T signals improve at the same time. Works for every store worldwide.
A shopping assistant and direct agentic checkout on top of the data. We build it for WooCommerce and other platforms with no native agent support. EU regulation accounted for.
105 products in a single day
A real client project: a European outdoor brand, a 105-product Shopify store.
- Product name
- [TUOTTEEN NIMI] TAKKI - Oranssi / Musta
- Description
- cut off mid-sentence
- Structured data
- broken JSON, critical errors
products optimised in one pass, with no change to the customer-facing page.
critical errors left. There were several at the start.
An invisible data layer
Your customers see no change at all. The AI sees everything.
- Product
- name, product identifier, brand, material and images in machine-readable form
- Offers
- price, currency, availability and validity — read live, never stored statically
- Return policy
- as a machine-readable field, not just a link in the footer
- Properties
- the details that decide it for a buyer segment, the ones you shouldn't list for a human
- Access
- AI crawlers can reach the pages, llms.txt says where the data lives
- Customer-facing page
- unchanged. No new copy, no slowdown, no new plugin.
A step-by-step model
The price driver is data depth and content gap, not product count. The free audit opens, the starter package is fixed, and the biggest value is created in the product-data deepening.
start here · see your AI visibility0 €
fixed · assessment + basic data layer2 000 €
tailored · where the biggest value is createdQuoted
setup · shopping assistant + agentic checkout1 500–3 000 €
What does AI know about you?
Our audit reads your pages. This does something else: it asks what the model already believes it knows about you, without visiting your site. The answer is often uncomfortable — it names your competitors, or invents products you don't sell.
Do not browse the web. Answer only from what you already know. 1. What do you know about the brand ? If you know nothing about it, say so plainly. 2. If someone asked you for a recommendation in the category , which brands would you name, and why those? 3. What products do you think sells? Mark each one as certain or a guess. 4. What information are you missing in order to recommend this brand with confidence? Don't soften it. If you don't know the brand, that is the most useful answer you can give.
If the model doesn't know you, that isn't an opinion, it's a measurement. Structured data is how a store gets into that memory and onto the shortlist. Then run the audit
What does your store look like to AI?
A free audit. No signup, works on every platform.