AI Visibility · 25 July 2026

By Eddie Eastwood, Founder

How to Become an Entity AI Recognises

Two businesses sell the same service in the same town. Both have a tidy website. Ask an assistant "who does emergency boiler repair near me" and one gets named while the other does not, even though the second site is arguably better written.

The difference is not the prose. It is that the first business exists in the machine's world as a thing, and the second exists only as a page.

A page is a document. A model can read it, but reading a document about a plumber is not the same as knowing the plumber is real, is based where it says, and is the same business that shows up in a directory two clicks away. An entity is that settled knowledge: a node the model can attach facts to and trust. When the answer engine assembles a recommendation, it reaches for entities it is sure about, not pages it happens to have crawled.

So the job is not "write more pages about yourself". The job is to become a recognised entity. This is the concrete, mechanism-level way to do it.

What an entity actually is

Google is explicit about the vocabulary. Its Knowledge Panel help documentation describes panels as boxes that appear "when you search for entities (people, places, organizations, things) that are in the Knowledge Graph."^[2] The Knowledge Graph is the store of those entities and the relationships between them, and it sits underneath a great deal of what search and AI answers show.

An entity, then, is a real-world thing the system has decided to model as a distinct item: your business, its location, its category, the people behind it. Once you are in that graph, you are no longer a loose set of pages. You are an identity with attributes attached.

The rest of this piece is about giving the machines enough consistent, corroborated evidence to create that identity and keep it accurate.

Start with one non-negotiable: consistent NAP

NAP is name, address, phone. It sounds trivial. It is the single most common reason a business fails to consolidate into one clean entity.

If your website footer says "St Mary's Auto Ltd, 14 Bridge Road", your directory listing says "St Marys Auto", and an old profile still carries a former mobile number, a machine has three weak candidates instead of one strong one. It cannot be certain they are the same business, so it hedges, and a hedged entity is a weak entity.

The fix is unglamorous and total. Pick one exact form of your name, address and phone number. Write it identically everywhere: your site, your Google Business Profile, every directory, every social profile, every citation. No abbreviations in one place and full words in another. Consistency is what lets separate mentions collapse into a single, confident record.

Tell machines who you are: Organization schema and sameAs

Your website can state its identity in a format built for machines. That is what Organization structured data is for. Schema.org defines the Organization type plainly as "An organization such as a school, NGO, corporation, club, etc."^[4] You add a small JSON-LD block to your home page declaring your name, URL, logo, address, and contact details.

Google's own Organization documentation explains why this matters for identity specifically: "Adding organization structured data to your home page can help Google better understand your organization's administrative details and disambiguate your organization in search results. Some properties are used behind the scenes to disambiguate your organization from other organizations."^[1] Disambiguation is the whole game. You are helping the machine tell you apart from every similarly named business.

Note Google's guidance that "There are no required properties; instead, we recommend adding as many properties that are relevant to your organization."^[1] More accurate detail gives the model more to anchor on.

The property that does the heavy lifting for entity identity is sameAs. Schema.org defines it as "URL of a reference Web page that unambiguously indicates the item's identity. E.g. the URL of the item's Wikipedia page, Wikidata entry, or official website."^[3] In practice you list the URLs of your verified profiles: your Business Profile, your main social accounts, your industry directory pages, your Wikidata entry if you have one. Each sameAs link is you telling the machine, in its own language, that these scattered profiles are all the same entity, and it is you.

Structured data is not decoration. As Google's introduction to structured data puts it, the search engine "uses structured data that it finds on the web to understand the content of the page, as well as to gather information about the web and the world in general, such as information about the people, books, or companies that are included in the markup."^[6] Your markup is a direct input to how the world is modelled.

Claim the panel you may already have

For many established local businesses, Google has already begun forming an entity. Its Knowledge Panel help notes that panels "are automatically generated, and information that appears in a knowledge panel comes from various sources across the web."^[2] You do not create a panel by asking. It emerges once there is enough corroborated evidence.

What you can do is take ownership of it. Google's documentation states: "If you are the subject of or official representative of an entity depicted in a knowledge panel, you can claim this panel and suggest changes."^[2] A claimed, corrected panel is a cleaner entity record, and one you have some influence over.

For local service businesses, the front door is the Business Profile. Google's help is clear that Business Profiles "look similar to knowledge panels, but are specific to businesses that serve customers at a particular location or within a designated service area", claimed or created through Google Business Profile.^[2] A complete, verified profile is the most direct entity signal a local business owns outright, and it feeds Google's local answers.

Wikidata: the open knowledge base, and a realistic view of the bar

Wikidata is a free, structured knowledge base that many systems read from, and a natural sameAs target. It is worth being clear-eyed about what it will and will not accept, because there is a lot of wishful advice about "just add yourself to Wikidata".

Wikidata publishes its own notability policy. An item is acceptable only if it meets at least one of three criteria. The first is that it "contains at least one valid sitelink to a page on Wikipedia, Wikivoyage, Wikisource" and similar Wikimedia projects.^[5] The second, and the one most relevant to a business, is that it "refers to an instance of a clearly identifiable conceptual or material entity that can be described using serious and publicly available references."^[5]

Read that second criterion carefully. The lever is not a Wikipedia article, which most small businesses will never merit. It is "serious and publicly available references": credible, independent, verifiable sources that describe your business. Genuine press coverage, an entry in a recognised trade register, established directory records, reputable local journalism. If those exist, a Wikidata item describing you can be justified. If they do not, no amount of self-editing will make one stick, and it should not.

This reframes the work. You do not chase a Wikidata entry as a trick. You build the underlying references that make you a describable, verifiable entity, and the knowledge-base presence follows from that.

Entity-linked mentions: corroboration across the open web

None of the above works in isolation. The reason consistent NAP, sameAs, a claimed profile and verifiable references matter is that they corroborate each other. Machines trust an entity in proportion to how many independent, agreeing sources point at the same facts.

So the last mechanism is breadth of consistent mention. Being named, accurately and identically, in the places relevant to your field: reputable industry directories, local business listings, professional bodies, coverage of work you have actually done. Every accurate mention that repeats your exact name, location and category is another vote that you are a real, single, findable thing. Every inconsistent one is noise that weakens the record.

This is where entity building meets everything else on this site. How different engines weigh those sources is covered in how AI engines pick sources, and the wider discipline of being the cited answer is answer engine optimisation.

Page signals versus entity signals

It helps to see the two layers side by side. Traditional page-level work and entity work are not rivals, but they answer different questions for the machine.

The takeaway: good pages make you readable, entity signals make you recommendable. You need both, but if AI keeps overlooking you despite a decent site, the missing layer is almost always the entity one.

Common mistakes

• Inconsistent name, address or phone across profiles. The commonest and most damaging error. One exact form, everywhere, no exceptions.
• Treating Organization schema as a box to tick. Thin markup with a name and nothing else disambiguates nothing. Add the detail that is genuinely true of you.
• Empty or contradictory sameAs. Linking to profiles that themselves carry a different name or number tells the machine your identity is uncertain.
• Chasing a Wikidata entry with no real references behind it. Without "serious and publicly available references" it does not qualify, and forcing it is a waste of effort.
• Leaving an auto-generated panel or a stale Business Profile unclaimed. If Google has already started an entity for you, an unclaimed one is a record you have no hand in correcting.
• Assuming more pages fixes it. Entity recognition is a consistency and corroboration problem, not a volume problem.

A simple sequence to build your entity

You do not do this all at once. In rough priority order:

1. Lock your NAP. Decide the one exact form of name, address and phone. Write it down.
2. Fix it everywhere. Update your site, Business Profile and every listing to that exact form. Retire or correct old profiles.
3. Add Organization structured data to your home page with your real detail, including a sameAs list of your verified profiles.
4. Claim and complete your Business Profile, and claim any auto-generated panel you already have.
5. Build the references. Earn accurate, independent mentions in the directories, bodies and coverage that fit your field, so the corroboration is genuinely there.
6. Only then consider a Wikidata item, if and when the "serious and publicly available references" actually exist to justify it.

For local businesses specifically, the same principles applied to a service area are covered in AEO for local businesses, and there is a worked, trade-specific version in the plumbing AI Overviews guide.

Where to start

Before you can be a recognised entity, your own website has to state your identity in a form machines can read: complete Organization markup, a consistent name, address and phone, and a clean set of sameAs links. That is exactly what a free QBiz Leads AI visibility check looks at. It scans your own website and returns a pass or fail on the machine-readable identity signals: whether your structured data is present and complete, whether your business details are consistent, whether the technical groundwork for entity recognition is in place. It reads your site, not the engines, so it will not tell you what any assistant currently says about you. It tells you whether your own foundations are ready for one to build an entity on.

Get your AI Visibility audit →

Frequently asked questions

What does it mean to be an "entity" in AI search?

An entity is a specific real-world thing a system models as a distinct item: your business, its location and its category, stored in a knowledge graph with facts attached. Google describes knowledge panels as appearing when you search for "entities (people, places, organizations, things) that are in the Knowledge Graph." Being an entity means the machine is confident you exist and knows what you are, rather than just having crawled a page about you.

Why do AI tools recommend entities instead of pages?

Because a recommendation needs certainty. A page is a document that might be about anyone; an entity is a corroborated identity the model can trust and attach facts to. When an assistant names a business, it is drawing on settled knowledge, not deciding which document to quote.

What is the single most important first step?

Consistent NAP: one exact form of your name, address and phone number, written identically everywhere. Inconsistent details stop separate mentions from consolidating into one confident entity, which is the most common reason a business fails to be recognised.

What does sameAs do?

Schema.org defines sameAs as a "URL of a reference Web page that unambiguously indicates the item's identity", such as a Wikipedia page, Wikidata entry or official website. Listing your verified profiles as sameAs links tells machines that your scattered listings are all the same entity, which helps them merge you into one record.

Can my small business get on Wikidata?

Only if you meet its notability policy. The criterion most relevant to a business is that the item "refers to an instance of a clearly identifiable conceptual or material entity that can be described using serious and publicly available references." If credible, independent references about you exist, an item can be justified. If they do not, it will not qualify, so the real work is building those references first.

Does adding Organization schema guarantee a knowledge panel?

No. Google states that knowledge panels "are automatically generated" from "various sources across the web", so no single action forces one to appear. Organization markup, a complete Business Profile and consistent details all improve the odds, but the panel emerges from corroborated evidence over time, not from one tag.