How AI recognizes your company: brand entity, Knowledge Graph and sameAs

What a brand entity is

In May 2012 Google launched the Knowledge Graph under the motto “things, not strings”: search began to understand that a query refers to a specific real-world object. At launch the graph held more than 500 million objects and more than 3.5 billion facts about them and the relationships between them. Google’s example: the query “taj mahal” may mean the monument, a musician, a casino or a nearby Indian restaurant, and the system has to work out which one is meant.

Companies are no different. “JW2D” is a string. The entity is a specific company with a legal name, registry code, address, website, social profiles, services and reviews. The more sources describe it the same way, the more confidently a system understands they all refer to one object.

Why this matters for AI

Google says AI Mode can tap into fresh, real-time sources such as the Knowledge Graph. Other AI assistants (ChatGPT, Perplexity, Claude, Copilot) do not publish how they store information about companies. But they all build answers from text and search results, so a simple rule applies: if sites describe your company differently or mix it up with another, AI may give the wrong address, attribute someone else’s reviews to you or not name you at all.

The risk is higher if the name is short, generic or shared with another company, product or person. Then it matters even more that every source links the name to the same facts.

What makes up a company entity

Organization markup and sameAs

Google recommends placing Organization markup on the home page or on a single page about the company, such as the About page. It helps Google understand the organization’s administrative details and tell it apart from others. There are no required properties; recommended ones include name, legalName, alternateName, url, logo, address, telephone, email, foundingDate, vatID, taxID and sameAs.

Google defines sameAs as the URL of a page on another website with additional information about your organization, and you can list several. What to include:

• official company profiles: LinkedIn, Facebook, Instagram, YouTube and others you actually maintain;
• your company’s page in the official register, if it has a permanent address;
• Wikidata and Wikipedia pages, if they exist;
• leave out: third-party articles about you, directories with incomplete data and long-abandoned profiles.

More on markup in schema markup for AI.

Google’s knowledge panel

A knowledge panel is the box on the right of Google results with information about a company. Google generates such panels automatically for entities in the Knowledge Graph, from various sources across the web, and updates them as that information changes.

• If a panel already exists, an official representative can claim it: click “Claim this knowledge panel” at the bottom of the panel and sign in to an official company profile such as Search Console or YouTube. After that you can suggest changes.
• If you are not a representative, you can report an error via the feedback link in the bottom right corner of the panel.
• For local businesses Google recommends Google Business Profile instead: it is meant for businesses with a location or service area.

Wikidata and Wikipedia

Wikidata is the open knowledge base of the Wikimedia projects. An item is acceptable if at least one condition holds: it has a page on Wikipedia or another Wikimedia project; it is a clearly identifiable entity described in serious, publicly available references; or it fulfills a structural need for other data. For a company the second condition usually applies, but only when independent coverage exists. On Wikipedia, its notability requirements and the rule against writing about yourself, see brand mentions for AI visibility.

How to check whether AI recognizes you

1. Search your company name on Google. Is there a knowledge panel or business listing, and is the data correct?
2. Ask AI directly: “What is [name], the company from [city]?” in ChatGPT, Gemini, Perplexity, Claude and Copilot. Compare the address, services, founding year and website with reality.
3. Check for confusion: does AI mix you up with a company or product with a similar name?
4. For technical staff: Google’s Knowledge Graph Search API shows whether the company exists in the graph as a separate entity. Google itself warns the API is not suitable for production-critical services, so treat it as a one-off check.

Short checklist

1. The same name, address, phone and description on the site, in listings and profiles.
2. An About page with legal name, registry code, founder and founding year.
3. Organization or LocalBusiness markup on the home or About page with a sameAs list.
4. A complete Google Business Profile or a claimed knowledge panel.
5. Social profiles linking to the site with the same description.
6. Independent mentions; Wikidata and Wikipedia once the company is notable.
7. A monthly check of what AI says about the company.

At JW2D, cleaning up the company entity is one of the first steps of AI search optimization: we find inconsistencies in company data and fix them on the site, in markup and in profiles.