Google and AI tools like ChatGPT no longer just match the words on your website. They try to work out who you are: what your business is, what it does, where it works and whether it can be trusted.
Entity optimisation is how you make sure they get that answer right. It sounds technical, but most of it is plain common sense, and almost none of it needs you to touch code.
What is entity optimisation?
Entity optimisation is the practice of making sure Google and AI engines clearly understand who your business is, what it does and how it connects to the topics you talk about.
An entity is the real-world thing behind the words: a person, a business, a product, a place or an idea. Search engines used to match strings of text. Now they try to match meaning.
For example, when a page mentions the iPhone, Tim Cook and macOS, Google knows “Apple” means the technology company, not the fruit. Nobody had to spell that out. Google worked it out from the company the word keeps.
Search Engine Journal describes entity optimisation as building connections between concepts in Google’s Knowledge Graph and removing ambiguity from your brand’s digital footprint (SEJ, 2026). In plain terms, you’re helping the machines stop guessing about you.
The engine looks for pages that contain the text “apple”.
- Apple crumble recipe for autumn
- Hawke’s Bay apple orchards open for picking
- Apple releases a new iPhone
- Is an apple a day really good for you?
Same word, four different things. The engine can’t tell which one you meant.
The engine reads the surrounding clues first:
- iPhone
- Tim Cook
- macOS
One clear thing, with known facts and relationships. That’s an entity.
Entities vs keywords
Keywords are the words people type. Entities are the things those words refer to. Here’s how the two compare.
| Aspect | Keyword SEO | Entity optimisation |
|---|---|---|
| Focuses on | The words people type | The things those words refer to |
| How engines read it | Matching text | Understanding meaning and context |
| Unit of work | Individual pages | Your whole business and its topics |
| Shows up in | Classic blue-link rankings | Rankings, Knowledge Panels and AI answers |
| The question it answers | “Does this page mention the term?” | “Is this the right business to trust?” |
Here’s a local example…Someone typing “plumber Tauranga” is using keywords. But when Google decides whether Able Plumbing is a real, trustworthy plumbing business that actually serves Tauranga, it’s thinking in entities: a specific organisation, a service, a place, and the links between them.
Keyword research is still useful. It tells you what people are looking for. Entities make sure engines recognise you as the answer. The two work together.
What counts as an entity?
- Person: the owner or a named expert
- Organisation: the business itself
- Product or service: what you sell or do
- Place: your shop, office or service area
- Concept: a topic your business should own
A Devonport cafe isn’t just “a cafe”. It’s an organisation (the business), in a place (Devonport, Auckland), run by a person (the owner and head barista), selling products (single-origin coffee, gluten-free baking), connected to a concept (specialty coffee).
A Hamilton accounting firm is an organisation with people (two chartered accountants and their credentials), services (GST returns, payroll, business structure advice), a place (the Waikato) and a concept it wants to own (small business tax).
Able Plumbing is an organisation led by a person (a licensed plumber), offering services (hot water cylinders, blocked drains, emergency call-outs) across a place (the Bay of Plenty), linked to a concept (residential plumbing).
The example businesses are illustrative.
Why does this matter? Search engines and AI systems keep something like a file on every entity they know. Your job is to make sure your file is complete, accurate and consistent.
Why entity optimisation matters now
Nearly 60% of Google searches now end without a click. The answer appears on the search result page, and the searcher moves on.
of brands already appear in AI-powered search results (Goodfirms)
of marketers actually track their AI citations (Goodfirms)
growth in AI traffic to US retail sites, Oct 2024 to May 2026 (Semrush)
People increasingly get their answer from an AI summary before they click anything. That AI decides which businesses to mention, and it leans on sources it can identify and trust, not on the pages that repeat a keyword most often.
For a business owner, the point is simple. If AI systems can’t confidently identify your business and what it’s good at, you are less likely to be recommended.
| What changed | What it means | What to do |
|---|---|---|
| Answers appear before clicks | Fewer visitors from the same rankings | Be the source the answer cites |
| AI builds answers from sources it trusts | Being findable isn’t enough; you must be identifiable | Make your facts consistent everywhere |
| Each AI tool has its own favourite sources | Your website alone won’t cover you | Keep listings, reviews and profiles current |
AI search has changed the rules of visibility
Google’s AI Overviews, ChatGPT, Gemini and Perplexity don’t hand you ten blue links. They write an answer, then cite the handful of sources they trust most.
Entity signals, meaning consistent facts, clear topics and credible mentions, are what get your business recommended and linked from AI responses.
How Google and AI engines see your business
Google keeps an enormous map of things and the relationships between them, and modern AI systems use similar maps to check their answers. On that map, ideas link together in chains:
- Artificial intelligence
- Large language models
- ChatGPT
- OpenAI
Each link makes the others easier to understand. Your business deserves the same clearly connected spot. When your website, social profiles, directories and mentions all describe you the same way, engines see one business. When they disagree, engines go with whichever version is easiest to trust, and that’s often not the one you’d choose.
The Knowledge Graph
The Knowledge Graph is Google’s database of things and the connections between them. Each entity is a dot, and each relationship is a line.
You’ve seen it at work: the panel on the right when you search for a well-known brand or person comes from this graph. It also works quietly in the background, shaping which content Google and AI engines treat as authoritative, even when no panel appears.
The business at the centre. Every other dot strengthens Google’s confidence about what this business is, who runs it and where it works.
- Able Plumbing (organisation): the business at the centre. Every other dot strengthens Google’s confidence about what this business is, who runs it and where it works.
- The owner (person): a named, licensed plumber with a bio on the About page. Real people with real credentials make a business easier to trust.
- Bay of Plenty (place): the service area. When the same region appears on the website, Google Business Profile and directories, engines know exactly where the business works.
- Google Business Profile (listing): its name, address and phone number should match the website character for character.
- Reviews (third-party mentions): outside confirmation. They carry more weight than anything the business says about itself.
- Hot water cylinders (service): a dedicated, detailed page makes this a clear part of the business’s identity.
- Blocked drains (service): another core service that reinforces one clear topic, residential plumbing. Related services strengthen each other.
- Emergency call-outs (service): linking a service to a place helps engines answer searches like “emergency plumber near me”.
How AI engines decide who to recommend (and who to ignore)
Here’s a practical test you can run in five minutes:
- Google your exact business name. Note what appears alongside it: a panel, a map listing, reviews, social profiles.
- Ask ChatGPT and Perplexity “What does [your business] do, and where?” Compare the answers with reality.
- Ask each tool for “the best [your service] in [your town]”. Are you named? Who is?
Any differences between those answers point straight at gaps in your entity.
The encouraging part: Yext analysed 6.8 million AI citations across ChatGPT, Gemini and Perplexity and found that 86% came from sources brands already control, such as their own websites and business listings (Yext). Much of what AI says about you is something you can influence.
Each engine also has favourites. In the Yext data, Gemini leaned more on websites while OpenAI’s models leaned more on listings. So your website, your listings and your reviews all need to tell the same story.
What entity optimisation actually involves
The whole strategy fits into four pillars. Each one is manageable for a busy owner, and none require you to build anything technical yourself. Work through them roughly in order: get your facts and structure clear first, then build authority and mentions on top.
Keep your facts consistent everywhere
- Business name (exact spelling)
- Logo
- Address and phone number
- Website address
- Opening hours
- Your one-line description
- Your website
- Google Business Profile
- Social media
- Directories
- Review platforms
- Anywhere you’re mentioned
Contradictory facts blur your entity, and engines then struggle to know which business is the real one.
Semrush points to Patagonia as a good example: its AI visibility stayed strong partly because outside sources such as REI, OutdoorGearLab and Reddit described the brand consistently.
Your action: start with your Google Business Profile, because Google and many AI tools draw on it heavily. Then make your website match it exactly.
Structured data without the jargon
Structured data (often called schema markup) is a set of labels added behind the scenes of your website. They tell engines exactly what each thing is: a business, a person, a service, a page. The labels come from a shared vocabulary at schema.org.
You don’t need to write code. Popular website builders and WordPress SEO plugins add the main labels for you, and most developers can set up the rest in an afternoon.
| Label | What it tells engines | Example |
|---|---|---|
Organization | This is a business, and these are its official details | Name, logo, website, social profiles |
LocalBusiness | This business serves customers in a place | Address, hours, service area |
Person | This real person is connected to the business | Owner’s name, role, qualifications |
Service | This is something the business offers | Hot water cylinder replacement |
An honest caveat: structured data is self-declared, so engines trust it less than outside confirmation. It supports your entity. It doesn’t create it.
Your action: ask your developer or agency what structured data your site has now, and to add Organization, Person and LocalBusiness labels where they’re missing.
Build authority on a few topics, not many
Engines judge what you’re about from the pattern across all your content. For an accounting firm, pages about tax, payroll and business structure reinforce each other. A one-off post about bicycle maintenance just blurs the picture.
Practitioners on Reddit keep landing on the same conclusion: entity work only pays off when it’s built around a clear topic and backed up by consistent coverage.
- Pick the three topics your business should be known for.
- Map every service page, blog post and social post to one of those three.
- Fill the gaps in those topics before branching into anything new.
Your action: write your three topics down and check every new piece of content against them before it’s published.
Earn mentions and citations that AI trusts
Reviews, press coverage, directory listings and being referenced by industry sites all reinforce your entity. Semrush and Yext both found that AI descriptions of brands draw on reviews, community discussions and independent publishers, not only on the brand’s own pages.
We saw the same thing in our own study of NZ life insurance answers: insurers’ own websites earned only 16% of the citations, while comparison and adviser sites took 43%.
What others say about you counts for more than what you say about yourself.
A small weekly habit- Reply to every new review, good or bad.
- Check that one profile or listing is still current.
- Ask one happy customer or partner to mention what you do, in their own words.
Your action: book a recurring 20-minute slot each week for the three tasks above.
The 7-step entity optimisation checklist for non-technical owners
Every step below is something you can either do yourself or hand off in one short conversation. Tick them off as you go.
0 of 7 steps done
Be patient: results build over months, not days.
How to know it’s working
- A Knowledge Panel for your business appears in Google, or an existing one fills out.
- AI tools name your business when asked about your industry or area.
- Your non-branded pages start ranking for relevant topics.
- Citations in AI answers start pointing to your website.
You don’t need expensive tracking tools. Most brands already appear in AI search without ever tracking it, but you can track the basics for free. Repeat the searches from step 1 each month, save the screenshots side by side, and watch the picture move from fuzzy to sharp.
Frequently asked questions
Is entity optimisation the same as SEO?
Not quite. Traditional SEO optimises pages for keywords. Entity optimisation makes sure search engines and AI recognise who you are, what you do and how your content connects. It’s part of modern SEO, and the two work together. See entities vs keywords.
Does entity optimisation replace keyword research?
No. Keywords still tell you what people search for, and entities describe what those searches mean. Keyword research guides your content; entity work makes that content recognisable and citeable.
Is SEO dead now that AI answers questions?
No, but it has changed. Fewer searches end in a click, so being cited inside AI answers matters more. A strong entity and solid content pull traffic from both classic search and AI tools. Our guide to ranking in AI search goes deeper.
I don’t use schema markup. Is that a problem?
Structured data helps, but it’s only one signal. Consistent information across the web, real reviews and focused content matter more. Add the labels when you can, and don’t wait for them before doing everything else. See the four pillars.
How long does entity optimisation take to show results?
Expect several months. Engines need time to digest consistent signals, and AI systems update on different schedules. Visible results typically arrive somewhere between month 3 and month 12 of sustained consistency.
Start with ten minutes today
Search your business name in Google, ChatGPT and Perplexity, and screenshot the results. It’s free, it takes minutes, and it shows you exactly where your entity stands right now.
Go to the checklistSources & further reading
- Goodfirms: SEO Statistics 2026: AI Search, Rankings & Zero-Click Trends, Apr 2026
- Yext Research: analysis of 6.8 million AI citations across ChatGPT, Gemini and Perplexity, Oct 2025
- Semrush: 2026 AI Visibility Index (126 million prompts), Jun 2026
- Aggarwal et al.: GEO: Generative Engine Optimization, Princeton University (KDD 2024)
- Search Engine Journal: What does “entity optimization” mean?, Jun 2026
- Relevance & Intent: NZ life insurance AI visibility study (ongoing), 25 September 2026 snapshot.
