New surface · ABR v0.1.0
Can an AI not only find your product, but correctly operate your storefront?
Shopify enabled WebMCP across Liquid storefronts, so agents can now search your catalog, read products, manage the cart, progress to checkout and answer policy questions inside the buyer’s real browser session — with no app to install. That is a second surface, separate from UCP’s server-to-server protocol layer, and it fails in different places.
Two surfaces, scored separately
UCP
Protocol readiness
Server-to-server. Can an agent platform ingest your catalog, resolve variants, trust your policies and delegate a purchase? This is the seven-category v0.1 score.
WebMCP
Agent Browser Readiness
In-session. Does the agent’s tool call actually work on your theme, does the cart keep the configuration, and does checkout progress? Good UCP does not imply this.
| Browser category | Weight |
|---|---|
| Tool exposure & discoveryWhich WebMCP storefront tools the store exposes, and whether an agent can discover and describe them. | 20 |
| Product & variant resolutionWhether in-session search and product tools resolve to the correct product and the correct variant or configuration. | 20 |
| Cart action fidelityWhether add-to-cart preserves quantity, variant and LINE attributes carrying per-item configuration, plus cart-level attributes for order-wide metadata. | 25 |
| Checkout progressionWhether the agent can progress to checkout without blockers, with truthful totals, shipping and taxes. | 20 |
| Policy & support answerabilityWhether shipping, returns, warranty and FAQ questions are answerable by the agent in session. | 15 |
Six metrics, not one pass/fail
Every eval case records these independently, so you can see exactly where the revenue leaks.
Discovery success
The agent surfaced a candidate product at all.
Correct product
The candidate was the product the buyer actually asked for.
Correct variant / configuration
Size, material, finish or configuration matched the request.
Browser-agent action success
The in-session tool call executed instead of silently failing.
Cart fidelity
The cart still carried the full configuration as line attributes, with cart-level attributes reserved for order-wide metadata.
Checkout readiness
The agent could progress to checkout with correct totals and shipping.
Live Lab: the gap is the diagnosis
Relationship neutrality
Independent public-web reference. Inclusion does not state or imply customer status, partnership, endorsement, participation, authorization, or private access.
Named profiles use only independently observable public-web information. No private, API, account or relationship data appears anywhere in these outputs.
| Store | UCP | Agent Browser |
|---|---|---|
| Zinus MéxicoBenchmark / positive control | 70 | 61 |
| Photobook MéxicoStress test / complexity control | 53 | 50 |
Both readings are provisional and public-web only: WebMCP support today runs through a Chromium origin trial, so in-session results stay pending until evals are executed in an authorized session. Commerce4 does not guarantee ranking, recommendation, placement or revenue on any AI platform.