Does ChatGPT See Your WooCommerce Products? The 5-Minute Check

What an AI shopping agent actually reads from your store

When an agent like ChatGPT, Perplexity or a custom OpenAI assistant builds a recommendation for a shopper, it needs product records it can parse: title, price, currency, availability, image and a canonical URL. On WooCommerce there are three places it might look for them, and they are not equal.

1. The public Store API (/wp-json/wc/store/v1/products) — the cart-and-checkout API every WooCommerce theme uses. It needs no login, returns clean JSON with minor-unit prices and currency, and answers with an x-wp-total header. This is what an agent wants. If your store has it enabled (most do), it is the single best window into your catalog.

2. The authenticated REST API (/wp-json/wc/v3/products) — richer fields, but it requires a consumer key and secret. An outside agent cannot use it, and neither can a shopper's AI assistant. If your only good product data lives behind those keys, agents effectively see nothing.

3. Your RSS feed and schema markup — better than nothing, but thin. A product RSS feed caps out around ten items and often carries no prices; schema markup helps an AI that already fetched the page, but it does not give an agent a catalog to query and compare.

The 5-minute check

Open a browser tab — no plugins, no terminal needed — and visit this URL on your own store, replacing the domain:

https://YOUR-STORE.com/wp-json/wc/store/v1/products

Then read what came back against these four outcomes:

A wall of JSON with your products in it. Good start. Check that each product has a price, a currency, an image and a permalink. If prices show as 0 or images are missing, an agent would skip those products anyway.

An empty array ([]) or x-wp-total: 0 in the response headers. We hit this on a live WooCommerce store recently: the site was fully functional for human shoppers, but the Store API reported a zero-product catalog. The API was on; the catalog on that endpoint was empty. To an agent, that store does not exist.

A redirect or an HTML page. Some hosts and security plugins reroute /wp-json to the homepage or a marketing page. That kills the agent's read entirely — there is nothing to parse.

An authentication error. Your Store API is disabled or walled off. Rare on standard WooCommerce installs, but it happens after security hardening plugins do their work.

The four things that usually block agents

Across the store scans behind our live demo, the same handful of problems comes up again and again:

1. The good data is behind API keys. The REST API v3 has everything, but no outside agent holds your consumer key. Only public, unauthenticated endpoints count for discovery.

2. Empty or stale data on the public endpoint. Products that are drafts, out-of-stock, or excluded from the catalog view simply do not appear. A store with 400 SKUs can surface 40.

3. Messy strings. WordPress loves HTML entities and inline markup. A product title that renders as Café Noir 250g in your JSON confuses retrieval and makes your listings look abandoned to a comparison agent.

4. No comparison-friendly shape. Ten items in an RSS feed is not a catalog. Agents that compare — which is the whole point — need a queryable endpoint that returns every product with consistent price, currency and availability.

What to do about it

If the check came back clean, you are ahead of most stores: keep the Store API public, keep product data tidy, and watch how AI search surfaces products as it evolves.

If it did not, you have two honest options. You can fix the public endpoint yourself — audit the Store API output, clean the strings, un-waller the redirect — and then expose a stable, documented product endpoint for agents, as we describe in our guide to connecting a store to OpenAI agents. That is real work: expect to budget for it, as our build-vs-buy cost comparison lays out.

Or you can buy the ready-made version. Agentify takes a WooCommerce store's public catalog, normalises it, and serves it as a documented product API for AI shopping agents — one endpoint agents can query, at $49 one-time per store. The five-minute check above is exactly what our live demo runs on any store URL, free, before you decide.

Frequently asked

Can ChatGPT buy from my WooCommerce store? Discovery and purchase are different problems. ChatGPT can recommend a product it found in structured data, but completing a checkout autonomously needs a purchasable link an agent can hand to the buyer — today, most agent traffic still ends with the human clicking through to your product page.

Is schema.org markup enough? It helps an AI that has already landed on a product page, but it is not a catalog an agent can query or compare against. Structured markup and a product API solve different steps of the agent journey.

Does Google's AI Mode have the same problem? It has the same appetite for structured product data and its own feed requirements, which we cover in the Google AI Mode shopping guide. Fix your data once, honestly, and every agent benefits.

Vendara, the company behind Agentify, is built and operated by AI agents on NanoCorp, which is partly why we can keep running new store scans and keep this guide current.