WooCommerce

WooCommerce is reached through the same WordPress connection, so there is no second setup: connect WordPress and RankX AI detects whether WooCommerce is present.

What it does is deliberately narrow. It reads your catalogue, flags products too thin to sell or rank, and rewrites descriptions and SEO fields. It changes no price, no stock level, no status and no product URL, and it is given no input for any of them.

Setting it up

Nothing beyond the WordPress connection. When RankX AI probes what your site supports, it reports whether WooCommerce is present, and the product tools appear if it is.

The WordPress account you connected as decides what is possible: RankX AI acts as that user and can never do more than that user could.

What it reads

The catalogue, with prices, stock status, and how much description each product carries.

That last one is the useful column. A product with a two-line description is a product an AI shopping surface has almost nothing to work with, and RankX AI flags the ones that are too thin to sell or rank rather than making you scan the list.

Reads are gated behind their own scope on the MCP surface rather than the general read scope, because they return prices and stock for a whole catalogue using your stored credential. See MCP authentication.

What it can change

Product descriptions, and the SEO title and meta description, subject to the same plugin constraints as any other page. See the WordPress integration.

Every change dry-runs first, showing you the exact before and after, and applying is a second explicit step.

What it deliberately cannot change

Permanent limits rather than a roadmap:

• No price. The tool is given no price input, and the price is reported before and after so you can confirm it did not move.
• No stock level.
• No status. A draft stays a draft, a published product stays published.
• No product URL or slug.
• No orders, customers or refunds. Those are reads only, and by construction: no write function exists to call.

Creating a product

wp_create_product adds a new product and always creates it as a draft. It cannot publish, so a human reviews the price and details in WooCommerce and publishes there. Publishing makes a product buyable, and that is not a decision to automate.

A price is required on creation, because WooCommerce treats a missing price as free.

Calling it twice creates two products. Unlike content creation, there is no idempotency key here, so never retry a creation whose outcome you are unsure of. List the catalogue and look.

Products have no undo

The one thing to know before changing anything.

WooCommerce keeps no revision history for products. A post or page can be rolled back to a WordPress revision; a product cannot, and RankX AI reports that plainly rather than letting you find out afterwards.

So for products specifically: read the dry run properly, and keep your own copy of a description you might want back.

Where this connects to AI Shopping

AI Shopping measures whether your products are recommended when a shopper asks an AI assistant, by matching the product cards a surface showed against the catalogue RankX AI holds.

Two links between the two features are worth making explicit:

An incomplete catalogue reads as a loss. A product RankX AI does not hold cannot be matched, so an answer recommending it is recorded as recommending somebody else. Your visibility reads lower and the merchant list reads longer. See products and the catalogue.

Thin descriptions are a common reason a product is not selected. These surfaces read listings, and the WooCommerce read is what tells you which of yours are too thin. That makes the two features a loop: measure with AI Shopping, fix with WooCommerce, measure again.

Driving this from an assistant

Three tools, behind their own scope:

"List WooCommerce products whose description is too thin to sell, then rewrite the worst three. Dry run first, and confirm the prices did not change."

See the WordPress tool reference.

Where to go next

• The WordPress integration, for the connection.
• AI Shopping, for what the catalogue is measured against.
• Products and the catalogue, for the matching.