Skip to content
Guide contents

User guide3 min

AI readiness

A score out of 100 for how easily an AI answer engine can fetch, understand and quote your site. It is built from checks the crawler already runs, with each weight shown.

How to get there#

Explore → AI readiness.

  1. 1

    Open AI readiness in the sidebar

    After Performance.

  2. 2

    Or press the AI readiness count on the site overview

    It is one of the four figures in the band at the top.

app.pixyscan.com/w/…/s/…/ai-readiness

The AI readiness screen: a readiness ring out of 100 with its band word, four counts beside it, and beneath it a card per weighted group with its pass rate and check count.
The score at the top, then the groups it is made of, each with its weight.

The three groups#

Each carries a different weight in the score.

FieldWeightWhat it does
Answer signals×70Whether your writing suits its audience, whether a page delivers what its title promises, whether AI crawlers are allowed to fetch it, and whether there is enough text to quote.
Machine-readable meaning×20Whether your pages state what they are about in a form a machine can read, mostly structured data.
Document structure×10Whether the page has a heading structure a machine can follow.

A group that did not run is left out of the calculation rather than scored as zero, and the counts beside the ring show how much was actually measured. A partial scan therefore gives a fair score rather than a low one.

Lower on the same screen

The AI readiness work list: failing checks in rank order with severity, reach and points back, and a footer noting the points belong to the health score.
Failing AI checks, in order. The points shown are health-score points, not points off this score.

llms.txt is graded on its contents#

Having the file is not enough if its links are broken.

Both llms.txt and llms-full.txt are fetched and checked: does the file follow the format, and do the links inside it still work. Which pages belong in the file is up to you, so coverage is only a suggestion and does not affect the score.

It does not predict whether you will be cited

The score measures things you control: access, structure and clarity. Whether a model actually quotes you depends on the model. Treat it as a checklist, not a forecast.

If the score is low because crawlers are blocked, fix it in Crawlability → robots.txt, not here.

If the product does not match this page, the page is wrong and we would like to know. Tell us