AI and search engines use JSON-LD schema to understand who you are and cite you correctly. Check what structured data your page exposes and whether your Organization entity is defined.
Measuring rankings in
Runs the full audit on your live site. Your report opens in your dashboard.
This extracts every JSON-LD block on the page, parses each one, and lists the schema types it declares. Blocks that fail to parse are reported separately from blocks that are absent, because they are different problems with different fixes. It then looks specifically for an Organization entity, which is the node that says who published this, and for the sameAs property that ties that entity to your profiles elsewhere.
Schema is the only part of a page written for machines rather than people, and it is how a search engine or an assistant resolves ambiguity. Without an Organization entity a model reading your page knows the words on it but has no confirmed identity to attach them to, which is exactly the condition under which it credits a directory or a competitor instead of you. sameAs is what closes that gap: it states that the company on this page and the company on that LinkedIn profile are one entity.
No JSON-LD at all is a failure and the most common result on small sites. An invalid block is worse than no block in one specific way: it is silently ignored, so you get the confidence of having done the work and none of the benefit. Start with Organization and WebSite on every page, then add the type that matches the page itself, Product, Article or FAQPage. Do not mark up content that is not visible to a reader.
Structured data is one check of thirty-six and it is worth four points of eighty-seven, which is roughly how much it should matter. The audit puts that number next to the ones that outrank it, so you can see whether schema is genuinely your gap or the thing you were about to do instead of the gap.
One trailing comma, in a block a CMS field assembled from a list of social profiles.
The comma after the last array item. JSON has no tolerance for it, the parser throws, and the entire block is discarded rather than partially read. Nothing warns you: the page renders identically, the markup looks right in view-source, and every downstream system behaves exactly as though you had never added schema at all. The same failure comes from a smart quote pasted out of a document and from an unescaped quote inside a string. This is why the report separates absent from present and unparseable — the second one has already cost somebody an afternoon.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Organization",
"name": "Example Co",
"sameAs": [
"https://www.linkedin.com/company/example",
]
}
</script>Organization and WebSite site wide, then the type that matches the page. Article on posts, Product on product pages, FAQPage where you genuinely answer questions on the page. Breadcrumbs are cheap and help on deep pages.
Not directly. It changes how a result is displayed and how confidently a system can identify your entity, both of which affect clicks and citations. Treating it as a ranking lever leads to marking up things that should not be marked up.
JSON-LD. Google recommends it, it sits in a script tag instead of being tangled through your markup, and it survives template changes that would otherwise break inline attributes.
Written up properly, for the question a result usually raises next.
More free tools in ai visibility, and the rest of the set at all free tools.
Free account, no card. Paste your URL and get a real, scored report of your AI and search visibility.
Measuring rankings in