Make your store visible to AI agents, and keep the checkout yours.

Shoppers are starting to ask ChatGPT and Google to find and buy for them. I make WooCommerce stores readable and buyable by those agents: structured catalog, Stripe’s Agentic Commerce Protocol, Google Merchant Center, the WooCommerce MCP server, and semantic search, with the transaction staying on your store.

  • Free 30-min call
  • No obligation
  • Reply within 1 business day

A good fit if

Sound familiar?

“Your Shopify competitors already show up inside ChatGPT. Your WooCommerce store doesn’t.”

“Your product data lives in titles and descriptions nobody structured; an agent can’t tell your sizes from your colours.”

“You’ve read about ACP, UCP, MCP and the Abilities API and don’t know which of them you actually need.”

“You want the discovery agents bring without handing them your checkout, your margin or your customer data.”

What I do

The work

Catalog structure

Attributes, variations, GTINs, stock, shipping and returns made explicit and machine-readable: schema.org Product and Offer, clean feeds, consistent taxonomy. The foundation every protocol reads from.

Protocol setup

WooCommerce MCP server and the WordPress Abilities API so assistants can read products and orders; Stripe for WooCommerce and the Agentic Commerce Protocol for checkout through agents; Google for WooCommerce and Merchant Center for Google’s AI shopping. Only the ones that apply to your store, configured and tested.

Semantic search on site

Search that understands how shoppers describe products, so the same catalog quality that helps agents helps your own visitors. I built AI Search for this.

Checkout and data boundaries

Where the transaction happens, what data leaves your store, and what agents are allowed to do, decided and documented before anything goes live. Agents get a catalog, not your customer database.

Testing with real agents

Scripted runs with ChatGPT, Claude and Google against the live store: find a product, compare, add to cart, check out. Fixed until they succeed, then monitored.

How it works

From first call to shipped work

Assessment, one week

Catalog quality, current structured data, payments and feeds reviewed. Which protocols apply, in what order, with a fixed price.

Build, two to four weeks

Catalog work first, then protocols, then search, each tested with real agents on staging before going live.

Monitor

Agent runs repeated monthly, feed and schema health checks, and a short runbook for your team.

Two to four weeks after the assessment, fixed price. Fixed price for defined work, hourly for ongoing engineering. How I work with US teams.

Proof

“Samuel was fantastic, very responsive, even over a weekend and across opposite time zones. He has helped me manage all issues within the project, answering any and all of my questions. Would definitely recommend.”

Client via Codeable, 4/5

Questions

Questions I get on the first call

Which protocols matter right now?

The WordPress Abilities API and WooCommerce MCP are live or in beta; Stripe’s ACP and Google’s UCP are rolling out. The catalog work is the same for all of them, which is why it comes first. I set up what is live today and keep the store ready for the rest.

Do we lose control of checkout?

No. The models that are winning keep the transaction on the merchant’s store; the agent handles discovery. I configure it that way and document what each agent can and cannot do.

We’re on Shopify. Does this apply?

The thinking does; the setup is different and Shopify ships much of it natively. I work on WooCommerce stores, and on migrations in either direction.

What does it cost?

The assessment is a small fixed fee. The build is fixed price after that, typically a few weeks of work depending on catalog size and how many protocols apply.

Tell me about the project.

Thirty minutes, free, no pitch. You’ll leave knowing what it would take.

Or email hello@samuelsilva.pt