Skip to content
Platform / Improve

Entity audit for AI search

The Axiom GEO entity audit crawls your key pages and rebuilds the entity graph your site publishes through its structured data. It checks for stable @ids, connected schema nodes, name variants, verifiable identifiers and authors set up as Person entities. You get scored findings, each with what was found, why it matters and the fix.

public version at /entity-audit
Free
public version at /entity-audit
core entity signals it rebuilds and scores
5
core entity signals it rebuilds and scores
to every finding, what, why and the fix
3 parts
to every finding, what, why and the fix

Key points

  • It rebuilds the picture of your business that your structured data gives to machines.
  • It checks stable @ids, connected nodes, name variants, verifiable identifiers and authors as Person entities.
  • Every finding is scored and comes with what was found, why it matters and how to fix it.
  • A free public version of the same audit runs at /entity-audit.
  • The schema generator in the platform writes the missing markup once you know the gaps.

What does an entity audit tell you?

It tells you how your business looks to a machine that reads your structured data. Not the words on the page, but the identity your markup builds when all the pages are put together. Is it one clear business with a name, a location, proof of who it is and real people behind the content? Or is it a loose set of fragments?

Most sites fall somewhere in between. A theme adds one organisation block, a plugin adds another with a slightly different name, and blog posts credit "admin" as the author. None of that breaks the site. It does make it harder for search engines and AI systems to be sure about you.

How does it work?

The audit follows the same steps each time, so results can be compared from one run to the next.

  1. Crawl your key pages. It reads the pages that say the most about who you are, rather than every page on the site.
  2. Collect the structured data. Every schema node on those pages is gathered in one place.
  3. Rebuild the entity graph. It joins the nodes up the way a machine would, following the @ids that link them.
  4. Check the signals. It looks at stable @ids, connected nodes, name variants, verifiable identifiers and authors.
  5. Score and explain. Each finding gets a score, a note on what was found, why it matters and the fix.

What does it check?

The table sets out the five main signals and what good looks like for each.

SignalWhat good looks likeWhat goes wrong
Stable @idsEvery page points to the same @id for your organisationEach page creates its own copy of the business
Connected nodesPages, services and people link back to the organisationNodes sit alone with nothing tying them together
Name variantsTrading name and legal name are both stated and linkedThree spellings of the name with no link between them
Verifiable identifiersLinks to official records and profiles that prove who you areNothing a machine can check against
Authors as Person entitiesReal named authors with their own Person markupContent credited to "admin" or to no one

If the terms are new, our glossary entries on entities and the knowledge graph explain them in a few paragraphs each.

What do you see?

You get a list of scored findings. Each one says what the audit found on your pages, why it matters for search and AI, and the fix. The score helps you decide what to tackle first. The explanation helps when you need to pass the work to a developer, or to explain to a client why it is worth doing.

A finding might say that your organisation appears under two names with no link between them. It would then explain that machines may treat those as two businesses, and show how to state both names on one node.

Where the fix is new markup, the schema generator can write it. Where the fix is about authors, it often links to the author bio and Person schema that come with content packs.

AI engines have to decide which business a question is about before they can say anything useful. If there are two firms with similar names, or your own site describes you three different ways, the chance of a mix-up goes up. A clear entity graph lowers that chance.

It also helps with trust. Named authors with a clear link to the business support the experience and expertise signals discussed in our guide to E-E-A-T for AI search. None of this guarantees a mention in ChatGPT or an AI Overview. Nobody can promise that. It does take away one of the easier reasons to be left out or described wrongly.

There is a link to answer accuracy too. When AI engines get your details wrong, the cause is often outside your site. Sometimes it is your own markup saying two things at once. The answer accuracy checks show where engines are wrong, and the entity audit helps rule your own site in or out.

Can I try it for free?

Yes. The free entity audit runs on your key pages and returns the same kind of scored findings. It is a sensible first step if you want to see where you stand before looking at the full platform.

Who uses it?

SEO consultants use it at the start of a new engagement, as a fast way to show a client where the gaps are. In-house teams use it after a site rebuild to check nothing got lost. Agencies run it across client sites to find the ones that need structured data work first. Our longer guide on entity SEO for AI search covers the thinking behind it.

Where it fits

The entity audit is the site-wide view of your structured data. Page optimisation goes page by page and scores each one. The schema generator writes the fixes. Answer accuracy shows whether AI engines describe you correctly once the work is done.

Frequently asked questions

What is an entity audit?

An entity audit checks how clearly your website describes your business as a single thing that machines can recognise. It looks at the structured data on your key pages and whether it joins up, names you the same way, links to proof of who you are and credits real authors. Gaps make it harder for search engines and AI to be sure about you.

Is the entity audit free?

Yes, there is a free public version at /entity-audit. You enter your site and it runs the audit on your key pages. The same audit sits inside the Axiom GEO platform, next to the schema generator and page analysis, so you can move straight from a finding to the fix.

Why do entities matter for AI search?

AI systems and search engines build their understanding around entities, meaning people, businesses, places and things. If your site gives a clear, consistent account of who you are, there is less room for mix-ups with similar names. It does not guarantee a mention, but it removes a common source of confusion.

What is a stable @id and why does it matter?

A stable @id is a fixed web address used as the identity of a node in your structured data, such as your organisation. When every page refers to the same @id, machines can tell it is one business. If each page invents its own, you look like several loosely related businesses.

How is an entity audit different from a normal SEO audit?

A normal SEO audit checks pages one at a time for things like titles, links and speed. An entity audit looks across pages at the identity your markup builds. It asks whether the pieces connect into one clear business with named authors and proof of who you are.

How often should I run an entity audit?

A good rhythm is after any site rebuild, theme change or new plugin, since those often change structured data without anyone noticing. Beyond that, a check every few months is usually enough for a stable site. Running it before starting schema work gives you a clear list of what to fix first.
Reviewed
Entity audit for AI search: check how machines see you