Is your site built to be cited by AI?

What the report covers

Ten checks, all of them mechanical.

Nothing here is a model's opinion. Every check is a fetch, a parse and a count, which means every number in your report is one you can reproduce.

1. 01Crawler accessWhether robots.txt lets the answer engines read you at all. Most sites pass this. The ones that do not are usually blocking by accident.
2. 02llms.txtWhether you publish a plain-text map of what you do and where the good pages are.
3. 03SchemaWhether your structured data describes the business, or only says that a web page exists.
4. 04Raw versus renderedHow much of your page is in the HTML the server sends, before any JavaScript runs. Most crawlers never run it.
5. 05Answer-first structureHow many characters it takes to get from the heading to a plain answer, across your top pages.
6. 06Section independenceWhether a section can be quoted on its own, or leans on "as mentioned above" to make sense.
7. 07Citation densityOutbound links, dates and named sources per thousand words on your content pages.
8. 08FreshnessWhether pages carry dates, and whether those dates have moved this year.
9. 09Entity consistencyWhether your name, category and claims are spelled the same way everywhere a model might read them.
10. 10Page coverageWhich of the pages buyers ask questions about you have not written yet.

Where most sites land

We ran the first four checks on 152 software sites.

Before building this we wanted to know whether the checks separate anyone at all. These are the results across 152 reachable company sites, using the same four checks that run on screen.

Run in September 2026 against 152 B2B software company sites, every one reachable at the time of the run. Sites that did not respond were dropped rather than counted as failures. Our own site passes all four, which we mention because you were about to ask.