Search stopped being about matching keywords a long time ago. Google announced its Knowledge Graph in May 2012 under the line "things, not strings", and launched it with more than 500 million objects on file. Those objects are entities: people, brands and places, each with facts attached and links to other entities. To get recommended, you first have to exist as one the machine recognises and trusts.

So how do you make AI trust your brand? You make it easy for the machine to be certain about you. Describe your business identically everywhere. Get listed in the structured sources engines use to define entities. Earn mentions from sites with a reputation of their own. Then cover your topic so thoroughly that the connection between your name and your subject is obvious.

Most businesses have never thought about this. They have a website, a few profiles set up years ago by different people, a directory listing with an old phone number, and a LinkedIn page describing a service they stopped offering. A human reads that as mild untidiness. A machine reads it as two or three different businesses.

What an entity is (and why keywords are not enough)

A keyword is a string of letters. An entity is a thing. "Apple" the keyword could mean fruit or a company. Apple the entity is one specific organisation with a founder, a headquarters and a product range, and each of those facts is joined to other things in a graph.

Think about what has to happen when someone asks an AI assistant "who is the best accountant for tradies in Newcastle." Matching those words against your homepage will not produce a recommendation. The assistant needs businesses it can confirm are accountants, that operate in Newcastle, that work with tradespeople, and that have enough independent backing to be worth naming. If nothing on the web confidently ties your name to those attributes, you are not a candidate, however good your service is.

Keywords made you findable. Entity status decides whether you get recommended. We unpack the wider mechanics in our guide to how AEO and GEO sit alongside SEO, but the practical work starts here.

Be the same thing everywhere

Consistency is how a machine works out that you and "that company" are one entity. Google says as much about its own local results: businesses with complete and accurate info are more likely to show up. Every contradiction splits your identity a little further.

It sounds basic, and most businesses still fail it. "Identical" means:

  • One exact business name. You cannot be "Smith Plumbing" on the site, "Smith Plumbing Services Pty Ltd" on Google and "Smith's Plumbing" on Facebook. Pick the version customers use.
  • One core description. Write two sentences on who you are, what you do and who you do it for. Paste them into every profile that has an "about" field.
  • One set of contact details. Same address format, same phone number, same website URL. An old mobile number on a 2019 directory listing is a contradiction somebody has to resolve.
  • One list of services. If your site says you do five things and your directory listing says eight, the machine has to guess which is true.

Get into the sources that define entities

  • Knowledge bases. Google says the facts in its Knowledge Graph come from public sources, licensed data and content owners. A complete Google Business Profile and LinkedIn page are the two every business can fill in. Crunchbase and Wikidata suit some companies and not others.
  • Credible mentions. A reference from a publication or industry body with its own reputation is evidence you did not write yourself.
  • Reviews. What customers say about you, and how many of them say it, shapes the description an assistant passes on.

A word on Wikidata. It is a community-edited database with its own notability policy. Unless you have a page on Wikipedia or a sister project, an item has to describe something clearly identifiable that can be backed by "serious and publicly available references". An entry that stretches the truth or reads like an ad can be nominated for deletion, so only pursue one when you have independent coverage to point to.

Reviews deserve their own attention, because they are one of the few entity signals you can grow every week. We go deeper on that in why AI reads your reviews before it recommends you.

Keywords made you findable. Becoming an entity makes you trustworthy.

Own your topic as well as your name

Trust builds when you cover a subject completely. A pillar page with every supporting angle linked to it shows you own the topic. Then add Organization structured data, which Google says helps it "disambiguate your organization in search results", so the connections are spelled out in a form software can read.

On a real site, that comes down to five jobs:

  • A pillar page for each core service that answers the big question in full: what it is, who it is for, how it works, what it costs, what to watch out for.
  • Supporting articles for the narrower questions a buyer asks along the way. Each one links back to the pillar.
  • Organization schema on your home page or about page, stating your official name, logo, address and phone number.
  • Person schema for founders and key experts. People are entities too, and a named expert attached to your business adds credibility.
  • sameAs links inside that schema pointing to your official profiles. Schema.org defines sameAs as a page that "unambiguously indicates the item's identity", which saves the machine guessing which LinkedIn page is yours.

Schema does not replace reputation. Its job is to turn "probably the same business" into "definitely the same business."

Build brand authority AI respects

A step-by-step entity cleanup you can start this week

  1. Write your canonical facts. Official name, two-sentence description, services, service area, address, phone, website, founders. Keep them in one document and treat it as the source of truth.
  2. List every place you exist online. Search your business name, then your phone number, then any old address. Write down each profile and directory entry that turns up.
  3. Fix the contradictions. Update every profile you control to match the canonical facts. Request corrections on the ones you do not control.
  4. Complete the big profiles properly. Google Business Profile needs categories, services, photos and opening hours. LinkedIn gets the same description. Add Crunchbase only if you are the kind of company that would reasonably be listed there.
  5. Add schema to your site. Organization first, Person where relevant, with sameAs pointing to the profiles you just cleaned up.
  6. Map your topic cluster. Pick your main service, write the pillar, list the ten questions buyers ask, and schedule the supporting pieces.
  7. Start earning mentions. Industry associations, supplier pages, local business roundups, podcast guest spots, useful answers in community forums. Independent sources are what turn your claims into corroborated facts.

A worked example: the physio clinic with three identities

Say you run a hypothetical physiotherapy clinic in Geelong. You rebranded two years ago. The website uses the new name. Google Business Profile still carries the old one. A health directory lists your former address. Your LinkedIn page describes you as a "wellness studio," which was a phase.

Ask an assistant about sports physios in Geelong and you are nowhere. The model has no opinion on your quality. It holds fragments of three businesses and cannot confidently connect any of them to "sports physiotherapy in Geelong."

Now run the cleanup. One name everywhere. One description that says exactly what you do: sports injury physiotherapy in Geelong for runners and footballers. Old directory entries corrected. Organization schema with sameAs links. A pillar page on sports injury rehab, with supporting articles on the specific injuries your patients ask about. A handful of mentions from the local running club and a sports association.

The clinic is exactly as good as it was before. What changed is that a machine can now state facts about it with confidence.

Common mistakes that keep you invisible

  • Leaving old profiles alive. A forgotten listing with outdated details keeps contradicting you. Update it or remove it.
  • Describing yourself differently on every platform. Clever variations feel fresh to you and read as confusion to a machine.
  • Relying on your own site alone. A brand that only talks about itself on its own domain has no corroboration.
  • Chasing volume over coverage. Fifty loosely related blog posts do less than ten that fully cover one subject and link together properly.
  • Treating schema as a magic switch. Markup clarifies what is already true. It cannot create authority the wider web does not support.

Run the audit

Ask an AI assistant to describe your brand and your category. If the answer is vague or wrong, you have found your entity gap. Closing it is some of the most valuable marketing work you can do this year.

Make it systematic. Put the same questions to ChatGPT, Gemini, Perplexity and Google's AI answers. Smith Plumbing would ask "What is Smith Plumbing?", then "Who are the best plumbers in Geelong?", then "What is Smith Plumbing known for?" Record the answers in a spreadsheet. Repeat monthly. Our guide to what AI is telling your customers about you walks through how to read the results.

Check the description for accuracy first. Then see if you get named in the category question, or only when someone asks for you by name. Last, note which competitors appear instead and which sources back them up. Those notes become your to-do list.

Make the machine certain about you

Entity work is not glamorous. It is a spreadsheet of profile logins, a block of JSON-LD on your home page and months of steady reputation building. Most of your competitors will skip it for that reason.

If you would rather have it done properly, our digital marketing team runs the audit, the cleanup, the schema and the content as one joined-up project, so the machine sees one clear brand instead of a pile of fragments.

Frequently asked questions

What does it mean to be an entity in search?

An entity is a distinct, identifiable thing, such as a person or a business, that search engines store with defined attributes and relationships. Google's Knowledge Graph has worked this way since 2012. Being an entity means the machine knows who you are, what you do and where you operate, which is what lets an assistant recommend you with confidence.

How do I get ChatGPT to mention my business?

You cannot submit your business for a mention. You earn it by being easy to understand and well corroborated. Keep your name and description consistent everywhere and complete major profiles like Google Business Profile and LinkedIn. Earn mentions on credible independent sites and collect detailed reviews. Publish thorough content on your core topic, then check regularly what assistants say about you.

Does schema markup help AI understand my brand?

Yes, as a clarity tool. Organization and Person schema state your key facts in machine-readable form, and sameAs links connect your site to your official profiles so there is no doubt they belong to the same business. Schema will not create authority on its own. It removes ambiguity, which makes the trust you have earned elsewhere easier to recognise.

How long does it take to build entity authority?

Cleanup work, such as fixing inconsistent listings and adding schema, can be done in weeks. Earning independent mentions and reviews is ongoing and compounds over months. Changes also tend to show up sooner in assistants that search the live web than in those leaning on older training data. Measure monthly and stay consistent.