AI-referred shoppers convert better, spend more, and land closer to purchase. We give them the information they need to make the right decision.
+42%
AI conversion lift vs. non-AI, Q1 2026. One year earlier the same channel converted 38% worse.
Adobe Digital Insights · Apr 2026
+393%
YoY AI-referred traffic to US retailers.
Adobe Digital Insights · Q1 2026
2.75×
more likely to land directly on a product page. Journey compression in production data.
Shopify Enterprise Blog
+49%
Shopify first-party: AI-referred sessions convert higher than organic search.
Risley, Shopify Enterprise Blog · May 2026
+14%
Higher AOV from AI-referred Shopify shoppers, with wins in 23 of 25 product categories.
Risley, Shopify Enterprise Blog · May 2026
The Field Test
We tested the endpoints to see just how many merchants were ready for agentic traffic.
What we discovered
Platform feeds and discovery protocols make products visible. They do not reliably carry why a product should be recommended: social proof, material, fit, color, promotion context, image meaning, or brand-specific ranking judgment.
2,483
Shopify storefronts
Top US-listed Shopify merchants, queried against the public UCP discovery surface with one identical signed request per merchant.
Sartorial field audit · May 2026
Across a live six-merchant UCP catalog test (query: “jacket”, limit 3), the discovery feed is uniform but the brand-truth signals agents need for comparison and recommendation are not.
UCP field audit, 2026-05-17: cuyana.com, everlane.com, gymshark.com, oliversweeney.com, taylorstitch.com, unrl.com.
| Signal | Cuyana | Everlane | Gymshark | Oliver Sweeney | Taylor Stitch | UNRL |
|---|---|---|---|---|---|---|
| Description chars (html) | 161 | 146 | 693 | 828 | 393 | 198 |
| Real compare-at price | No | Yes ($248) | No | Yes ($256) | Yes ($198) | Yes ($128) |
| Variant compare-at | No, all 3 | No, all 3 | No, all 3 | No, all 3 | No, all 3 | No, all 3 |
| Color as structured option | Yes | No (title) | No | Yes | No | Yes |
| Size option | Yes | Yes | Yes | Yes | Yes | Yes |
| Shopify taxonomy categories | Yes | Yes | No | Yes | Yes | Yes |
| Image alt text populated | Yes | Yes | No | No | Yes | No |
| Tags per product | 1 | 10 | 55 | 8 | 21 | 18 |
| Tag namespacing | flat | key: value | key:value | Key_Value | Key: Value | merch flags |
| Useful tag prefixes | none | fabric, category | feature, colour, activity, fit, season | Colour, type, group | Color, Material, Construction | None* |
| Variant-level media | Yes | Yes | Yes | Yes | Yes | Yes |
| Reviews / ratings | No | No | No | No | No | No |
* UNRL: public tags are mostly internal merchandising flags such as campaign, report, and cohort labels. They are unlikely to mean anything useful to LLMs or agentic shoppers.
Merchants are missing the brand-truth layer
Platform feeds and discovery protocols make products visible. They do not encode merchant-specific judgment: when to recommend, substitute, avoid, warn, upsell, route, or explain a product in the brand’s voice.
Sartorial
Publishing into AI surfaces and merchant-authored reasoning: fit, substitution, policy, routing, audience.
Shopify + protocols
Product feed exposure, UCP-style discovery, platform-neutral catalog surfaces.
Hosted platforms
Catalog, inventory, pricing, checkout, storefront, orders.
Shopper asks an AI surface
Shopify catalog (product feed exposure)
Sartorial · merchant-authored brand truth
Shopify checkout (PDP, payment)
Qualified recommendation, purchase
Early access
Sartorial is invite-only while we onboard the first wave of brands. Drop your email and we’ll be in touch with pilot details.
No spam. Unsubscribe anytime.