Structured vs. Traditional: A Live Comparison
We publish a real product catalogue two ways — as a typical HTML storefront would, and as our Agentic Commerce approach does — then show the verifiable difference. No projections below; every figure is checked against the live sites. Last verified 13 September 2026.
The same business, published two ways
The left card is how most e-commerce catalogues still work. The right card is our live proof-of-concept, built the way we build for clients.
Typical E-Commerce Catalogue
- Product data exists only inside rendered HTML — an AI agent has to scrape and guess at price, stock, and specs.
- No standardized machine-to-machine checkout path; an agent can't verify or transact without a human in the loop.
- Optimized for keyword matching, not for exposing structured facts a model can trust.
- Every redesign risks silently breaking whatever an agent was previously able to parse.
Mobile Fun AI Storefront (Live Demo)
- All 500 products publish Schema.org
Product,ItemList, andFAQPageJSON-LD — verified against the Schema.org validator. - The full catalogue is queryable as structured JSON at
/api/v1/ucp/catalog— no scraping, no guessing. - A machine-readable manifest at
/.well-known/ucpdeclares catalog and checkout capabilities to any agent that asks. - Price, stock, ratings, and FAQs are typed fields, not prose an LLM has to interpret.
What we actually verified
Checked directly against the live sites on 13 September 2026 — not estimated.
Structured data is valid
Every product's Schema.org JSON-LD passes validation, and the UCP manifest at /.well-known/ucp resolves with a valid catalog and checkout endpoint declared.
Catalogue is fully machine-readable
All 500 products across 11 categories are queryable via one JSON API call — the same data a human sees on the browse page, exposed as typed fields.
Search exclusion is intentional, not accidental
The demo sends X-Robots-Tag: noindex on every response — confirmed live. It's deliberately kept out of Google/Bing so it never competes with a client's own site; its only job is agent discovery.
What we can't claim yet
We'd rather tell you what's still open than round up.
Classic search ranking
Both this site and the demo are new enough that neither currently ranks for anything in Google or Bing — that's a wash today, and it isn't the metric that matters here anyway, since the demo is permanently excluded from classic search by design.
AI-answer citation share
Tracking how often an AI catalogue actually gets cited by AI Overviews or shopping agents, head-to-head against an unstructured competitor, needs a live analytics connection we haven't provisioned on this demo yet. We'll publish real numbers here once that's live — not before.
See it yourself
Browse the live demo, query the catalog API directly, or talk to us about doing this for your own catalogue.