Published
From page to entity
Search systems have spent a decade moving from matching strings of text to identifying entities: the specific organisation, person or place that a page refers to. Generative answers depend on that identification completely, because an answer names a company rather than linking to a document and hoping the reader works it out.
To name a company confidently, a system needs a consistent account of what it is, where it operates, what it sells and who it serves. It assembles that account from the company's own site, from directories and profiles, and from whatever third parties have written.
Where the sources disagree, confidence falls. A low-confidence entity is one the system is less likely to volunteer, and more likely to describe vaguely when it does.
The ordinary ways companies contradict themselves
The inconsistencies are rarely dramatic, which is why they survive. A company calls itself a branding studio on its website, a digital agency on LinkedIn and a marketing consultancy in a directory listing. An office moves and three profiles are updated out of five. A trading name and a registered name are used interchangeably, along with an abbreviation nobody has documented.
Individually these are trivial. Collectively they describe three or four slightly different companies and no single one of them clearly.
The same applies to services. A page selling 'growth' where the buyer searches for 'performance marketing' is a translation problem for a human reader and an identification problem for a machine.
Google's guidance asks site owners to keep business information accurate across its own products for exactly this reason. The instruction is dull, which is why it is skipped in favour of more interesting work.
What a working About page contains
The version that does this job reads closer to a reference document than a manifesto, without being lifeless.
It states what the company does in plain language in the first paragraph, without metaphor. It names the services, the markets and the kinds of client served. It gives the location and a contact route. It names the people where they can be named, because individuals are entities too and their credibility attaches to the organisation.
It carries the evidence a third party might quote: when the company was founded, what it has done, what it specialises in, any verifiable credential or membership.
None of this prevents the page having a voice. It does prevent the page opening with three sentences of atmosphere before disclosing what the business actually is — a convention that costs nothing when a human reads it and costs identification when a machine does.
Structured data, in proportion
Organisation markup is worth adding, with a caveat. Google states plainly that structured data is not required for generative AI search and that no special schema exists for it.
What markup does is remove ambiguity cheaply. Stating the legal name, the address, the contact point and the profiles the company controls, in a format that admits no interpretation, costs an hour and eliminates a category of confusion.
It is not a ranking device and should not be sold as one. It is a way of saying the same thing everywhere without relying on a machine to infer it.
The cheapest audit in marketing
The exercise takes an afternoon and requires no tooling.
List every place the company is described: website, LinkedIn, Google Business Profile, industry directories, partner sites, old press coverage, conference bios. Write down, verbatim, how each describes the business in one sentence. Put the list side by side.
If those sentences do not agree, the machines are working from a contradiction, and no amount of on-page optimisation resolves it. Correcting them costs nothing but attention, and it is prerequisite work for every other tactic in AI-era search — including the expensive ones a company may already be paying for.
Common questions
Does my About page affect SEO?
Increasingly, yes. AI systems build a description of your company from your site, your profiles and third-party sources, and an About page that states plainly what the business does, where and for whom is the primary source they use.
Do I need Organisation schema markup?
It is useful but not required. Google states that structured data is not required for generative AI search. Markup's value is removing ambiguity about your legal name, location and contact details cheaply.
What is a brand entity in search?
The specific organisation a search system identifies behind the mentions of a name. Consistent descriptions across your site, profiles and third-party sources raise confidence in that identification; contradictory ones lower it.