43%*
serve AI search bots a block or captcha
Through the firewall or CDN, while a regular browser gets in.
NL/BE: 20%
RESEARCH ✦ UK EDITION ✦ NOVEMBER 2026
How AI-ready are
UK Magento stores?
I scanned 373 Magento stores in the United Kingdom with the same question I asked of Dutch and Belgian stores: can AI search engines such as ChatGPT, Perplexity and Google AI read, understand and recommend them? Here are the results, side by side, with all the caveats included.
MEDIAN SCORE · UK
67/100
- 0–390%
- 40–5931%
- 60–7961%
- 80–1009%
01 / AT A GLANCE
Six numbers,
two markets.
The overall scores are close. But UK stores turn AI search bots away twice as often, and their product data is less complete. Each card shows the UK figure, with the Dutch and Belgian figure underneath.
9%
score 80 points or more
Roughly one in eleven stores has the basics for AI search engines in good shape.
NL/BE: 11%
5%
block an AI search bot in robots.txt
Rare, but ten times as common as in the Netherlands and Belgium.
NL/BE: 0.5%
22%**
have brand, GTIN, price and availability complete
Exactly the data AI assistants use to compare products across stores.
NL/BE: 33%
29%**
only have older microdata, no JSON-LD
Typical of the default Luma theme. AI systems read JSON-LD more reliably.
NL/BE: 15%
9%
have an llms.txt
A new file for AI assistants. Not a standard yet, but quick to set up.
NL/BE: 8%
* Measured with simulated AI search bots; real bots may be treated differently. Treat this as an upper bound. Only a refusal (HTTP 403) or a challenge/captcha page counts; rate limits and connection errors do not.
** Among the 120 UK stores (and 114 NL/BE stores) where the scan found a product page automatically. Product fields were measured in JSON-LD; stores with microdata only count as incomplete there.
02 / FINDINGS
What is going on,
and what to do about it.
Three themes, each with the UK and NL/BE numbers, what they mean and what you can check yourself. Most of it can be fixed in a few hours.
01 / ACCESS
Twice as many stores turn AI search bots away
At 43% of UK stores, at least one of the three AI search bots I tested (from OpenAI, Perplexity and Anthropic) got a refusal or a captcha, while a regular browser from the same address got in without trouble. In the Netherlands and Belgium that was 20%.
It is not a matter of policy in robots.txt. Training bots, which collect content to train AI models, are blocked about as often in both markets (19% in the UK, 21% in NL/BE). The difference is in the firewall or CDN, which often stops AI bots by default. I did not measure which CDN or firewall each store uses, so I cannot say exactly why the UK is higher.
Caveat: I simulated the bots using their user agent, from the Netherlands. Real bots are sometimes recognised as a "verified bot" and let through, so treat this number as an upper bound. And some stores block bots on purpose; that is their right.
BLOCK OR CAPTCHA, PER BOT*
CHECK IT YOURSELF
- Open your store with the user agent of OAI-SearchBot, PerplexityBot or Claude-SearchBot and see whether you get an error or a captcha.
- Check which bot rules are active in your CDN or firewall, and distinguish between search bots and training bots.
- Make a deliberate choice: if you want to appear in AI answers, let the search bots through.
02 / PRODUCT DATA
More old microdata, fewer complete products
For 120 UK stores, the scan could find and analyse a product page automatically. 57% use JSON-LD, the format search engines prefer, against 74% in NL/BE. 29% only have the older microdata that the default Luma theme outputs, and 14% have no structured data at all.
That carries through to the details. A product code such as a GTIN or EAN is in the JSON-LD of 28% of UK stores, the brand in 44%. Only 22% have brand, GTIN, price and availability all complete, against 33% in NL/BE. Reviews are rare: 11%.
The good news: on 97% of UK product pages the price is right there in the HTML, readable without JavaScript (NL/BE: 88%). The content is there; it just is not labelled in a way AI assistants read best.
FIELD PRESENT IN JSON-LD
CHECK IT YOURSELF
- Run a product page through Google's Rich Results Test and see which fields are in your Product schema.
- Still on microdata? Add JSON-LD for products, and map the brand and EAN attributes you already have in Magento.
- Do you have reviews? Put them in the data too: only 11% of UK stores do so now.
03 / DISCOVERABILITY
The basics are similar on both sides
On discoverability the two markets are close. 90% of UK stores have meta descriptions (NL/BE: 94%) and 69% have a sitemap that can be found, slightly more than in NL/BE (66%). Organization data, which tells who is behind the store, is less common: 47% against 58%.
About one in eleven UK stores already has an llms.txt (9%). That file is not an official standard yet and there is no guarantee it helps, but it is quick to set up.
% OF STORES
CHECK IT YOURSELF
- Add the link to your sitemap to robots.txt, so every crawler can find it.
- Add Organization schema with your company name, logo, contact details and social profiles.
- Consider an llms.txt with a short description of your store and links to your main categories.
THE BOTTOM LINE
AI assistants recommend what they can reach, read and trust.
03 / METHOD & LIMITATIONS
How I
measured it.
Every store in both markets got exactly the same automated scan, on the same day, from the same location. That makes the markets comparable, but it is still a snapshot with limitations. I list them openly.
01→
Selection
From a public ranking of the world's most visited websites, I took all .uk domains (such as .co.uk): 30,500 in total. The NL/BE edition used the same ranking.
02→
Detection
Of those, 415 were automatically recognised as Magento stores, based on technical traits in the HTML.
03→
Scan
Every store got the same checks as the free Shortlist scan. 373 UK scans succeeded, scanned on 12 November 2026 alongside 427 stores in NL/BE.
04✦
Anonymous
I only publish totals. No store names appear in this research, and that will not change.
LIMITATIONS
What to keep in mind
- This is a snapshot from November 2026. Stores change all the time.
- The ranking mostly contains larger, high-traffic stores. Smaller stores may look different. UK stores on a .com domain are not included.
- AI bots were simulated using their user agent, from the Netherlands. Real bots, or visitors from the UK, may be treated differently.
- Product data was only measured for the 120 stores where a product page was found automatically, and product fields only in JSON-LD.
- The score comes from my own scan method. It measures whether AI assistants can read your store, not how often they actually mention it.
The Dutch and Belgian results, with more detail per finding, are in the NL/BE edition.
Spot something that is off, or missing? Let me know. This is the first edition, and I want to repeat and improve the research next year.
04 / YOUR STORE
How does
your store score?
Run the same scan on your own store for free. Within a minute you see your score and the main points to improve. Or send me a message if you want to talk through the results, or have feedback on the research. I work remotely with stores across the UK and Europe.
Run the free AI scanTELL ME MORE
Thank you for your message!
I read every message myself and will get back to you as soon as possible.