Tool · product data
Will an AI agent buy from your store?
A shopping agent never looks at your page, it reads a product record. Enter your store address and we will check your product pages and show which fields are missing before a purchase can close. The measurement is deterministic: we read structured data rather than guessing with a language model.
- No signup
- No email required
- Result in seconds
How we measure
We fetch robots.txt, find your sitemap and sample a few product pages from it. From each one we read schema.org structured data and check the nine fields a purchase cannot close without.
We respect robots.txt, identify ourselves in the User-Agent header and space out requests. We do not store your store content, only the result of the field measurement.
This samples your catalogue, it does not audit it. A few pages reveal the template the whole store is built on, though individual products can deviate.
Identifier (SKU or GTIN)
Without it an agent cannot tell your product is the same item a competitor sells, so it cannot compare offers.
Price
The price has to be a number in the data, not text on an image or buried in a script.
Price currency
A number alone means nothing. With no currency an agent cannot compare your price to any other.
Availability as a machine value
„Only a few left” describes a mood. An agent needs a value it can check against a rule.
Selling unit and multiple
Piece or case pack. Without it an agent orders ten times too many, or too few.
Tax included or not
In B2B this decides the order total. An agent that does not know gets the sum wrong.
Delivery cost and date
Without it there is no final total, and the total is what decides which store wins.
Variant attributes as separate fields
When size lives inside the product name, an agent cannot narrow the choice and goes to a competitor.
Return terms as data
A link to your terms page will not do. An agent weighs purchase risk from data, not prose.
What we found in Polish stores
Measured on 25 August 2026 across twenty-four stores. Seven could be read, seventeen could not: seven carry no structured data on product pages, four block automated traffic, four have no sitemap listing products.
The last three fields never appeared once, and tax status showed up in a single store. The sample is small and skewed towards large chains with eCommerce teams, so these numbers are a floor for the problem rather than a ceiling.
| Price | 7 of 7 |
| Currency | 7 of 7 |
| Availability as a machine value | 5 of 7 |
| SKU or GTIN identifier | 4 of 7 |
| Variant attributes | 3 of 7 |
| Tax included or not | 1 of 7 |
| Selling unit | 0 of 7 |
| Delivery date | 0 of 7 |
| Return terms | 0 of 7 |
You have the gaps. What now?
Product data gaps are fixed once and then pay into search visibility, model answers and sales channels at the same time. If you want someone to run the numbers on your catalogue, start with a conversation about the platform and the data behind it.
Questions about the agent readiness check
- Does this tool use AI?
- No, and that is deliberate. The check reads schema.org structured data and tests for specific fields, so the result is repeatable and every gap can be pointed at in your page source. A language model here would only add uncertainty. The AI is on the other side: shopping agents read this data, and we check whether there is anything for them to read.
- Why did my store come back unreadable?
- Usually one of three reasons: no sitemap listing product pages, bot protection rejecting automated traffic, or product pages with no Product structured data. Treat it as a result rather than a failed test, because a shopping agent hits exactly the same wall.
- Do these gaps hurt me in Google?
- Partly. Some of them (price, availability, identifier) feed product results and Merchant Center. But this check covers a wider set, because closing a transaction needs more than displaying an offer. Selling unit and tax status change nothing about how your result looks in Google, yet they decide whether an agent totals the order correctly.
- You only check a few products. What does that say about my catalogue?
- Product pages come from one template, so the sample shows what that template publishes. If the selling unit is missing on four random pages, it is almost certainly missing on all of them. The exception is stores running several templates for different categories, and that is the only case where checking more pays off.
- Do you store my store data?
- We do not keep your page content. We keep only the field measurement, so that we can occasionally publish aggregate market statistics, always without naming individual stores.
