Agent-ready is about actions, not appearances.
A Commerce.com integration gives browser agents structured shopping tools. The useful work is defining what those tools may do.
What changed
Cloudflare describes a Commerce.com rollout that exposes shopping tools through WebMCP, with merchant opt-outs. Its example keeps the final checkout with the shopper. Chrome’s documentation still presents WebMCP as experimental.
Go to the source ↓Why it matters
A shop has two separate jobs: help a buyer understand the offer and let a permitted assistant perform a task correctly. A new description of the brand does not fix an ambiguous product variant, an out-of-date cart or a missing confirmation step.
The opportunity is to make a small set of actions explicit: finding a product, selecting a valid option and preparing a cart. Each action needs a defined result and a clear boundary before money changes hands.
What remains unproven
This is a supplier’s account of its own integration, not an independent completion-rate study. It establishes neither universal browser support nor increased sales or AI recommendations. Support must be checked for the actual commerce platform and customer workflow.
The useful next step
Map one shopping journey from search to checkout. Identify where a tool could help, which actions need confirmation and how the ordinary storefront remains usable when the agent path is unavailable. Validate that journey before a wider rollout.
Define the action, its limits and the human handoff.