Blog

We scanned 148 online stores — only one passed the AI-readability test

3 July 2026 · 3 min · Store Operators

ChatGPT, Perplexity and Google AI recommend products to buyers every day. We wanted to know how many stores are actually readable to them — so we scanned a national market's worth: 148 Finnish e-commerce sites.

Method

In early July 2026 we ran our AI Readiness scan against 148 e-commerce sites. The scanner reads a store the way AI agents do: it finds a product page, checks structured data (schema.org), product-data completeness, content quality and AI-crawler access. On 61 of the sites the scanner found an automatically analyzable product page — the per-check numbers below refer to those.

Anyone can run the same scan on any store for free, so the results are reproducible.

Results: average 64.5/100 — and a single A

Grade distribution across all 148 stores: 1 × A, 59 × B, 56 × C and 32 × D. Average score: 64.5 out of 100. In practice, almost no store in the sample is in shape for AI recommendations — which also means the head start is still up for grabs, in any market.

The starkest per-check numbers:

Failure Share of stores
Price and currency missing from machine-readable data 88.5%
Stock status missing 88.5%
Reviews not exposed in structured data 86.9%
Description doesn't answer buyer questions 83.6%
GTIN/EAN product identifier missing 78.7%
No concrete specs or materials in the description 60.7%
No schema.org Product markup at all 49.2%

Why this matters right now

In March 2026 Shopify started feeding store product data directly into ChatGPT, Gemini and Copilot. Distribution is automatic now — but AI picks products whose data is complete and precise. When 88% of stores don't even expose a price in machine-readable form, AI can't recommend them no matter how good the products are.

The same applies across platforms: WooCommerce outputs Product structured data by default, but brand and product identifiers are missing from it entirely without extra work — and those are exactly what AI uses to match your product to a buyer's question.

Buying behavior is already shifting: according to Adobe Analytics, visitors arriving from AI search convert 31% better than other traffic, and Shopify reports AI-driven orders growing 11–15× year over year.

What a merchant should do

  1. Measure your baseline. The free scan shows in 30 seconds what AI sees in your store — and what it doesn't.
  2. Fix the structured data. Price, availability, brand, GTIN and reviews in schema.org form on every product page. This is a technical change that doesn't alter your page design at all.
  3. Rewrite descriptions as facts. "A quality jacket" tells AI nothing. "Gore-Tex Pro, 320g, waterproof 28,000mm" tells it everything. Materials, dimensions and use cases decide.
  4. Verify crawler access. Good news: this was the rarest problem — only 2 stores out of 148 blocked AI crawlers.

We do the whole package at a fixed price in 48 hours, with a money-back guarantee. But even if you do the work yourself or with your own partner: do it soon. AI recommendations are being handed out now, and there's no queue yet.

Method note: scans run July 3, 2026 with our public AI Readiness scanner (structured data, product-data completeness, content quality, crawler access; weighted scoring). Sample: 148 Finnish e-commerce sites, of which 61 had an automatically analyzable product page. Per-check shares refer to those 61.

What does your store look like to AI?

Run the same scan on your own store — 30 seconds, no signup, works on every platform.

Scan your store free