Skip to main content
Analyse

Free tool

Can a generative system work out who you are?

Before anything can represent your business accurately, it has to be able to establish what the business is, what it offers and whether that can be confirmed anywhere else. This checks whether those signals are present on your home page.

Enter a website address. One page is requested, and nothing is stored.

The problem

Being readable is not the same as being understandable

A generative system answering a question about your sector does not browse your website the way a customer does. It retrieves passages, resolves the organisations mentioned in them to known entities, and assembles an answer from what it can confirm. Anything ambiguous tends to be dropped rather than guessed at, because a wrong specific is worse than a vague generality.

This is where a great many competent websites lose out. A page that reads beautifully to a human — evocative headline, confident tone, services implied through case studies — may never actually state, in so many words, what the organisation is and what it sells. A person infers it in seconds. A retrieval system has nothing to quote.

The signals below are the ones that close that gap. None of them is exotic and none requires a rebuild. Most are a matter of saying plainly what everyone internally already assumes is obvious.

What is checked

Three groups of signals

Each is checked for presence and legibility, not for quality — a tool can see that an organisation has been described, not whether the description is any good.

Who you are

Entity clarity

Whether the organisation is described as a named entity in structured data, whether it names itself consistently, whether it links to profiles it controls elsewhere, and whether any sentence on the page plainly defines what it is.

What you do

Service understanding

Whether what is offered appears in structured data, whether headings describe services rather than gesturing at outcomes, and whether service, about and contact pages are reachable from the home page at all.

Why you should be believed

Structure and authority

Whether expertise is attributed to identifiable people, whether the page's position in the site is declared, whether questions and answers are explicitly structured, and whether an llms.txt declaration exists.

Interpretation

How to read the result

Everything marked as worth reviewing is a signal a machine reader could not establish from this page. That is not automatically a fault. A consultancy that deliberately keeps its home page short and pushes detail to service pages is making a legitimate design choice, and the check has simply looked at the wrong page for that information.

What should concern you is a cluster. If entity clarity is uniformly missing — no organisation described, no external corroboration, no definitional sentence — then a system reading this site has genuinely nothing to anchor on, and any representation it produces will be assembled from elsewhere.

The single most valuable item on the list is usually the least technical one: a plain sentence stating what the organisation is and who it serves. It costs nothing, it is the thing most likely to be quoted, and it is missing from a remarkable proportion of otherwise excellent websites.

A useful definitional sentence

Names the organisation, states its category, states who it serves and where. “Harrowgate Legal is a commercial property practice acting for owner-managed businesses across Yorkshire.” Unremarkable to read, and directly quotable.

What it replaces

“We partner with ambitious businesses to unlock growth.” This states a posture rather than a fact. It could describe a law firm, an agency or a lender, which is precisely the problem.

What this cannot establish

  • Whether any generative system has ever retrieved or cited this site. That information is not published by any of them.
  • Whether the structured data is accurate. Presence and type are checked; whether every declared property matches the visible page is a manual review.
  • Whether the organisation has genuine authority on its topics, which is the part that decides whether being legible translates into being used.
  • What is on the rest of the site. Entity signals are frequently stronger on an about page than a home page.
  • Whether content rendered by JavaScript exists, since nothing is executed during this check.

Powered by Genyra

Questions about GEO

Does passing this mean ChatGPT will cite us?
No, and any tool that tells you otherwise is guessing. No generative system publishes its retrieval criteria, and several change substantially between releases. What this checks is whether the information such a system would need is actually present and legible. That is a precondition, not a guarantee — much as having a well-structured page is a precondition for ranking without being a promise of it.
Why does it only look at the home page?
Because the home page is where organisational identity is normally established, and because a free check that crawled entire sites would be an imposition on the sites being checked. The consequence is that this understates sites whose entity signals live on an about page. If several checks here come back unclear, look at your about page before concluding the signals are missing.
Is structured data the whole of GEO?
It is the most measurable part, which is why tools concentrate on it, and that has produced a widespread misconception. Structured data makes existing facts machine-readable. It cannot create authority, originality or accuracy, and marking up a page that says nothing specific simply makes the emptiness easier to parse. Structured data is necessary and nowhere near sufficient.
Should we publish an llms.txt file?
You can, and it costs almost nothing. It is an emerging convention with no confirmed support from any major generative system, so treat it as a low-cost bet rather than a strategy. Anyone presenting llms.txt as the answer to generative visibility is selling something.
How does this differ from a GEO audit?
This checks whether signals are present. An audit examines whether what those signals point at is worth retrieving — whether the organisation has said anything specific, original and verifiable, how it compares to the sources currently being drawn on for its topics, and what would need to be true for it to become one of them. Presence is the easy half.

Going further

IntelligenceGEO IntelligenceThe full module: what is assessed, how it is weighted and what it deliberately avoids claiming.ServiceGEO AuditA reviewed assessment of entity clarity, retrievability and citation readiness across the whole site.ReportGEO ReportThe deliverable: findings, comparison against the sources currently being drawn on, and a sequenced direction.

Legible is the starting point, not the destination

A GEO audit examines whether what your signals point at is actually worth retrieving — and what would need to change for it to be.

Request a GEO auditRun the full snapshot

GEO assessments evaluate observable discoverability, structure, clarity and authority signals. They do not guarantee visibility or citations within third-party AI systems.