How the score is built.
The score sums up how far an agent gets on your site: find, read, buy, complete. Every stage consists of checks from the outside. This page says which question each stage asks, what green means and what the score does not measure.
The question each stage asks.
Being found
The question
Does an agent find you at all?
Green means
An agent finds you without a detour.
Machine-readable
The question
Does it understand what you sell, what it costs and when it ships?
Green means
Product, price, delivery and returns are data to it, not guesses.
Shoppable
The question
Can it order from you without needing a human?
Green means
It finds the purchase path and can place an order.
End-to-end
The question
Does it get from the first search to the order confirmation without getting stuck?
Green means
The path runs through, without clicks on your website.
What we check on each stage in detail is in your report: the specific gap and the next step, per stage.
Where does an agent get stuck on your site?
We measured 50 shops from the outside. Median 35 out of 100. None had a purchase path of its own for agents.
Four stages · One score · Free
One score, four status words.
The score is a number from 0 to 100. Each stage carries a status word, green, yellow or red, always as a word, never just a colour. The seal starts at 80 with an active Watch; below that, or without Watch, the verify page says "no seal".
What the score does not measure
Revenue. Ranking. Mentions in AI answers. The last one is stage 1 and the field of our partners. The score measures whether an agent can read you and buy from you, not whether it recommends you.
From the outside, without a login.
IthildimBot/1.0 checks from the outside, the way an agent sees your site: it honors robots.txt, keeps to rate limits, identifies itself by name and never buys anything. How it works and how you lock it out is described under /bot.