We Checked 1,161 UK Shops: What AI Shopping Can't See

AI shopping assistants like ChatGPT, Google AI Mode and Perplexity recommend specific products from specific shops. To do that, they have to read those shops: the product names, the descriptions, the price and stock, the page summaries.

We wanted to know how readable ordinary independent shops actually are. So between 9 and 11 October 2026 we opened up 1,161 UK online stores, read 598,382 product listings and 28,823 product pages, and counted what was there and what wasn't.

The short version: only 24.1% of the stores passed all four basic checks below. The rest had at least one gap that makes it harder for AI to understand or trust their products. Most of those gaps are fixable in an afternoon or a few weeks, depending on the size of the catalogue.

The headline numbers

Across all 1,161 stores (620 WooCommerce, 541 Shopify):

• 19.1% had no product data AI can read on any product page we opened. Across all pages, 22.4% had none
• 38.8% had thin descriptions (under 150 characters) on at least 15% of their products. 18.4% had them on at least half. In total, 24.4% of all products we read were that thin
• 32.7% had at least one product named with an internal stock code; 5.3% had it on 15% or more of the catalogue
• 29.7% of product pages had no meta description (the short page summary search engines and AI tools use)
• 47.7% of WooCommerce products had no short description
• Only 24.1% of stores passed all four core checks: product data on every page, meta description on every page, no stock-code names, and fewer than 15% thin descriptions

How we measured it

We want these numbers to be useful, so here is exactly what we did and what the limits are.

The stores. 1,161 independent UK online shops (all UK-based according to our store database), running WooCommerce (620) or Shopify (541). They come from our own outreach list, which targets small and mid-sized shops with roughly 10 to 2,500 products and their own catalogue, and leaves out obvious dropshippers. The median store had about 450 products. This is not a random sample of every UK website. It's a fair picture of "independent UK shops that sell their own range", which is the group we care about.

The audit. For each store, an automated audit:

1. 1Read the product list from the platform's public product data (WooCommerce's Store API, Shopify's public product list), up to 1,000 products per store
2. 2Opened up to 25 product pages spread evenly across the catalogue, with a normal browser user agent
3. 3Looked at the HTML each page actually serves: is there Product data (Schema.org JSON-LD, microdata or RDFa)? Is there a meta description?

What counts as what:

• Thin description: under 150 characters of plain text, roughly one sentence
• Stock-code name: a product name that starts with an internal code, like "BX40 Black" or "SKU-2231 Shelf"
• Product data: a Schema.org Product block in the page, in any of the three common formats

Who's in, who's out. We only counted stores where the audit was reliable: the product list was readable and at least 20 product pages loaded without being blocked. 421 audited stores were dropped for that reason (blocked by a firewall, too few pages, or the list wasn't readable), plus earlier audits that didn't keep these counts. So if anything, the shops here are the more open and reachable ones.

What we didn't measure here. The audit also looks at image alt text, reviews, brands and barcodes (GTINs) on product pages, but those counts weren't kept in the summary we analysed, so we're not reporting them. We also didn't test whether ChatGPT actually recommends these shops; this is about what the shops make available to AI, not about rankings.

Finding 1: One in five shops hides its product data

Product data (often called structured data) is the machine-readable block that tells AI the product's name, price, currency and stock status without guessing from the page layout.

• WooCommerce: 22.6% of stores had none on any page we checked, 25.6% had it on fewer than half, and 60.3% had it everywhere
• Shopify: 15.2% had none on any page, 17.2% on fewer than half, and 79.9% had it everywhere

This is the one gap that's mostly not about effort. Both WooCommerce and most Shopify themes output basic product data by default. When it's missing, something in the setup (a theme, page builder, plugin or headless front end) usually removed or replaced it. One check in Google's free Rich Results Test tells you which group you're in.

Finding 2: A quarter of all products are described in one sentence

This was the most common gap overall. 24.4% of the 598,382 products we read had descriptions under 150 characters.

• WooCommerce: 44.2% of stores had at least 15% thin descriptions, 22.4% had at least half. 28.3% of all WooCommerce products were thin
• Shopify: 32.5% of stores had at least 15% thin descriptions, 13.9% had at least half. 19.3% of all Shopify products were thin

Why it matters: when a shopper asks a follow-up ("is it waterproof?", "will it fit a small dog?"), the assistant needs the answer in your text. One sentence rarely contains it.

And good technical setup doesn't save you here. 25.7% of stores had product data on every single page and still had thin descriptions on at least 15% of products. The data says "this is a £34 jacket, in stock". The words never say why anyone would want it.

Finding 3: Code names are common, but usually only a few products

A third of stores (32.7%) had at least one product named with an internal code: 40.5% on WooCommerce, 23.8% on Shopify. But only 5.3% of stores had it on 15% or more of the catalogue, and 3.6% of all products overall.

So for most shops this is a tidy-up, not a project: find the products that start with codes and give them real names. Those few products are often imported or wholesale lines, which can be exactly the ones that need help being found.

Finding 4: WooCommerce shops skip the page summary

This is where the platforms differ most.

• WooCommerce: 50.1% of product pages had no meta description, and in 39.0% of stores none of the pages we checked had one
• Shopify: only 6.3% of product pages were missing one

The reason is simple. Shopify fills in a meta description automatically from the product description. WooCommerce doesn't; it usually comes from an SEO plugin like Yoast or Rank Math, and many shops never set it up.

Also on WooCommerce: 47.7% of products had no short description, the summary that usually sits next to the price. In 42.6% of WooCommerce stores, at least half the products were missing it.

Finding 5: Shopify shops start ahead, WooCommerce shops have more to gain

We count a store as having at least one big gap if it hits any of these: 15% or more thin descriptions, 15% or more stock-code names, product data on fewer than half its pages (WooCommerce), or meta descriptions missing on 30% or more of pages (Shopify).

• WooCommerce: 60.5% of stores had at least one big gap
• Shopify: 35.7% did

And for all four core checks together (product data and meta description on every page, no code names, under 15% thin descriptions):

• WooCommerce: 12.3% passed
• Shopify: 37.7% passed

Shopify's defaults do a lot of quiet work. WooCommerce gives owners more control, which also means more things left unset.

What this means for your store

If you run an independent shop, the odds are good that at least one of these applies to you. The encouraging part is how ordinary the fixes are:

1. 1Check your product data. Paste one product URL into Google's Rich Results Test. If there's no "Product" result, look at your theme and plugins
2. 2Name products for people. Search your catalogue for names that start with codes and rename them to say what the product is
3. 3Write real descriptions, best sellers first. What it is, who it's for, what it's made of, how it's used
4. 4Fill in page summaries. On WooCommerce, set up meta descriptions in your SEO plugin and fill in short descriptions
5. 5Make sure AI crawlers can get in. Check robots.txt and your firewall settings

We wrote step-by-step versions for each platform: the WooCommerce guide and the Shopify guide. For the background on why this matters now, see what AI commerce is.

Where Pickastor comes in

We ran this audit because it's the problem Pickastor is built for, so here's an honest note on what it does and doesn't cover.

Pickastor works on the words: it rewrites product names, descriptions, short descriptions and tags, adds product attributes and the phrases shoppers use, and pushes the approved changes back to your WooCommerce, Shopify, PrestaShop, Magento or BigCommerce store. It also publishes an AI product feed with every product you optimize. That covers findings 2, 3 and the short descriptions in finding 4, plus SEO titles and meta descriptions on Shopify, PrestaShop, Magento and BigCommerce.

It doesn't add product data markup to your pages, change robots.txt or llms.txt, or edit image alt text, brands, reviews or barcodes. Finding 1 and WooCommerce meta descriptions are settings in your theme and plugins, and they're worth fixing whether or not you ever use a tool.

You can check your own store for free with the AI Score, no signup needed. If you want help with the rewriting, the first 3 products are free and the rest is a one-off payment by catalogue size; see pricing.

Using this data

Journalists, bloggers and researchers are welcome to quote these numbers. Please cite "Pickastor, live audit of 1,161 UK WooCommerce and Shopify stores, October 2026" and link to this page. If you need a breakdown we haven't shown here, get in touch.