Entity SEO: making your brand a known thing

Search engines used to match words on pages to words in queries. They now also maintain large bodies of knowledge about things: companies, people, products, places and how they relate. Google calls its version the Knowledge Graph. AI models absorb similar knowledge from the text they are trained on and the pages they retrieve. When someone asks "which bike shops in Leeds do e-bike repairs" or "who makes the best merino base layers for winter running", the answer depends partly on what these systems know about the businesses involved as entities.

Entity SEO is the work of making sure your business is recognised clearly and described accurately. It is less about rankings for particular pages and more about the system knowing who you are, what you do and what you are associated with.

How systems recognise an entity

Recognition comes from repeated, consistent information across sources the system trusts:

• Your own website: the about page, contact details, product and service pages, structured data.
• Official profiles: business listings, social profiles, app store pages, marketplaces.
• Authoritative third parties: news coverage, industry associations, government registers, reference sites, partner and supplier pages.
• Mentions in context: articles, reviews, forum posts and lists that name the business alongside the topics it is known for.

When these agree on the name, what the business does, where it is and who runs it, systems can connect them into one confident picture. When they disagree, or when there is too little, the picture is blurry, and assistants either leave the business out or describe it wrongly.

Step 1: write the canonical description

Write down, once, the facts that define your business, and use them everywhere:

Fact	Example for a hypothetical business
Official name	Northfield Cycle Works
One-sentence description	An independent bicycle repair and e-bike service workshop in Leeds, founded in 2014.
Category	Bicycle repair shop
Location and service area	Address; serves Leeds and surrounding towns
Key services or products	E-bike motor and battery diagnostics, wheel building, servicing
People	Founder's name and role
Official URLs	Website, Business Profile, social profiles

The one-sentence description is surprisingly important. It is the sentence most likely to be repeated by directories, journalists and AI systems. Make it specific and true, and use the same wording, or close to it, on your about page, profiles and press page.

Step 2: make your website the clearest source

• An about page that states the canonical facts plainly: what the business is, where, since when, who runs it, what it is known for.
• Contact details in text, consistent with listings.
• Organization or LocalBusiness structured data on the home page or about page, with name, URL, logo, contact information and sameAs links to official profiles. See structured data.
• Author and team pages for the people associated with the business, with Person markup where appropriate.
• Product and service pages that name your offerings consistently, so systems learn what you are associated with.

A minimal example of Organization markup:

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Northfield Cycle Works",
  "url": "https://northfieldcycleworks.example/",
  "logo": "https://northfieldcycleworks.example/logo-512.png",
  "description": "An independent bicycle repair and e-bike service workshop in Leeds, founded in 2014.",
  "sameAs": [
    "https://www.facebook.com/northfieldcycleworks.example",
    "https://www.instagram.com/northfieldcycleworks.example"
  ]
}

Only list profiles you control and keep active. sameAs links to dormant or unofficial profiles add confusion, not clarity.

Step 3: align the profiles

Go through every official profile and listing and align it with the canonical description: name, category, description, address, website, logo. Old descriptions from years ago, former names and outdated services are common. The citations guide covers the local listings side.

Some reference sources carry particular weight because many systems read them, such as major business directories, industry registers and, for notable organisations, open knowledge bases. Do not create or edit entries about your own business on volunteer-edited reference sites in ways that break their rules on conflicts of interest; it tends to backfire. Earn mentions that independent editors can cite instead.

Step 4: earn mentions in the right context

For AI systems in particular, what a brand is mentioned alongside matters. A bike workshop mentioned repeatedly in local news, cycling club pages and forum threads about e-bike repairs becomes associated with exactly that. Ways to build those associations legitimately:

• Digital PR with data or expertise in your niche; see digital PR.
• Expert commentary from named people at the business on topics they know.
• Partnerships and memberships that come with public listings.
• Useful resources that others reference.
• Customer reviews that describe what you did, in their own words.

Unlinked mentions still count for recognition. Links are a bonus, and the reclaiming links guide covers how to request them politely.

Step 5: check what the systems believe

• Search your brand name and look at what Google shows: a Knowledge Panel, a Business Profile, sitelinks, the description used.
• Ask several AI assistants "What is [business name]?" and "What does [business name] do?" and note errors.
• Ask brand-free questions in your category to see whether you are mentioned at all, following the method in measuring AI visibility.

Trace errors to their sources. Wrong descriptions usually come from somewhere: an old listing, an outdated article, a similarly named business. Fix what you can, and make your own pages state the correct facts unambiguously.

Products and people are entities too

The same logic applies below the brand level. A product with a distinctive name, described consistently on your site, by retailers and in reviews, becomes something systems can recognise and recommend by name. Give each product a stable name and a clear one-line description, keep specifications identical everywhere you publish them, and avoid renaming products casually; every rename splits the mentions between two names.

People matter as well. Founders, engineers, chefs or advisers who write, speak and are quoted under their own names build expertise that systems associate with them and, through them, with the business. Author pages with clear roles and links to their published work help connect the person to the organisation. This is also what the E-E-A-T guide describes from the quality side.

A worked example

When Northfield Cycle Works asks assistants about itself, one describes it as a bicycle shop in a different Northfield in another country, and another lists services it stopped offering in 2019. The cause is easy to trace: a common place name in the business name, and a popular directory listing that still carries the description written when the shop opened. The owner adds "Leeds" to the one-sentence description everywhere it appears, rewrites the about page around the current services, updates the directory listing, adds Organization markup with the correct address and profiles, and asks the local cycling club, which links to the shop, to update its outdated description. Over the following months, brand questions start returning the right city and current services more consistently.

Common problems

• Name collisions with another business or a common word. Use a consistent qualifier ("Northfield Cycle Works, Leeds") and make the category clear everywhere.
• Rebrands where old and new names coexist in sources. State the change on the about page ("formerly known as…") so systems can connect the two.
• Founder and company confusion when a well-known person and their business share a name. Separate pages and clear Person and Organization markup help.
• Multiple locations described inconsistently. One parent organisation with clearly described branches.

Why it matters now

Classic rankings depend heavily on pages and links. AI answers depend more on what systems know and can retrieve about entities. A business that is described consistently, associated with the right topics in many independent sources, and represented by a clear website, gives every one of those systems an easier job. It is slow work with no single switch, but it compounds, and it is very hard for competitors to copy quickly.