Entity SEO and the Knowledge Graph - Semantic Optimization
Since 2012, Google hasn't thought only in keywords - it thinks in entities: people, companies, products, places and the relationships between them. That year Google launched the Knowledge Graph under the banner "things, not strings". Entity SEO (semantic optimization) is optimizing for that understanding: making search engines and AI models know unambiguously who you are, what you do and why you're credible. It's the foundation of Google visibility and AI citations - with one important caveat about the role of structured data, which I'll come back to.
Google and AI models think in entities, not keywords. Entity SEO optimizes for that understanding: entity disambiguation, consistency and corroboration (via Wikidata and schema). The foundation of Google visibility and AI citations.
Why entities matter more than ever
The reason is simple: both Google and AI models reason in entities, not words. This shifts the center of gravity of all SEO. An Ahrefs study of 75,000 brands found that brand mentions correlate with AI search visibility 3× more strongly than links (0.664 vs 0.218) - and a mention only works when the machine can tie it to the right entity. In other words: if Google and AI don't unambiguously know who you are, even the best digital PR dissipates into noise. Entity SEO is the work that makes every mention "land in the right account" - your brand's, not a similarly named company's.
What an entity and the Knowledge Graph are
In Google's sense, an entity is "a thing or concept that is singular, unique, well-defined and distinguishable". It needn't be physical - a person, company, place, event, work, idea. The Knowledge Graph is a database of such entities and facts about their relationships (at launch in 2012, over 500M objects and 3.5B facts, drawn from sources including Freebase, Wikipedia and the CIA World Factbook). Each entity has a machine ID (MID - e.g. the `/m/...` format from the Freebase era, or the newer `/g/...`).
On top came natural-language understanding: BERT (2019, affecting ~1 in 10 queries) and MUM (2021) mean Google interprets context and meaning rather than matching exact strings.
/// ENTITY GRAPH: LINKED BY @id AND sameAs
* Define site-wide entities once and link them via @id; sameAs "glues" them to the Knowledge Graph entity (Wikidata = strongest signal).
The three pillars of Entity SEO
1. Disambiguation. The engine must tell you apart from others with the same name. It does this through context and co-occurrence (what appears alongside the name): "Jaguar" + engine/dealership is a different entity than "Jaguar" + predator/species. Consistent context and related entities point to the right meaning.
2. Consistency. The same brand data everywhere - on the site and in external profiles. Mismatches (different names, descriptions, contact details) blur the entity and lower Google's "confidence" in who you are.
3. Corroboration. The more credible, independent sources confirm the same facts, the stronger the entity. Wikidata is key here - it has softer criteria than Wikipedia (it's enough that the entity is "clearly identifiable" and described by a reliable source) and it's a direct Knowledge Graph source.
/// THE THREE PILLARS OF ENTITY SEO
How to implement it technically
- Schema.org with an `@id` graph. Define entities (`Person`, `Organization`/`ProfessionalService`, `WebSite`) once and link them via `@id` across the site. It's a machine-readable map of your identity. Details: Schema.org structured data.
- `sameAs` to authoritative profiles. schema.org defines `sameAs` as "the URL of a reference page that unambiguously indicates the item's identity" - Wikipedia, Wikidata, official profiles. These "glue" your site to the entity in the graph.
- A consistent author. Every post and page should point to the same author (`author` → the same `Person` via `@id`), with credentials; mark the About page with the `ProfilePage` type.
- Entity-building content. A rich, unambiguous "About" page (the "entity home") - it feeds entity understanding, and independent sources corroborate it.
{ "@context": "https://schema.org", "@graph": [ { "@type": "Person", "@id": "https://yourdomain.com/#person", "name": "First Last", "sameAs": ["https://www.wikidata.org/wiki/Q000", "https://linkedin.com/in/profile"] }, { "@type": "Organization", "@id": "https://yourdomain.com/#organization", "founder": { "@id": "https://yourdomain.com/#person" } } ]}
Entity audit - a checklist of where to start
Before you start "building an entity", check how it looks to machines today. A practical checklist:
| Element | Control question | Priority |
|---|---|---|
| Entity home | Do you have one rich "About" page as a source of truth? | High |
| Schema with @id | Are entities (Organization, Person, WebSite) defined and linked via @id? | High |
| sameAs | Do you link to Wikidata and authoritative profiles via sameAs? | High |
| NAP consistency | Are name, address and contact details identical everywhere? | Medium |
| Wikidata | Is there an item about your brand with a link back? | Medium |
| Consistent author | Do all posts point to the same author via @id? | Medium |
| Co-occurrence | Does content unambiguously tie the brand to its topic (context)? | High |
Rule: first put your own site in order (entity home + schema + consistency), then secure external corroboration (Wikidata, mentions). The reverse order wastes signals - external mentions have nothing to "stick" to.
How to build a strong entity - step by step
Building an entity is a sequence, not a single action. Order matters, because each step reinforces the previous one:
- 1.Define the entity home. A single, rich "About" page that unambiguously describes who you are, what you do, for whom, and why you're credible (experience, work, data). This is your source of truth.
- 2.Build the schema graph. Define `Organization`/`Person`/`WebSite` once, link them via `@id`, and attach profiles via `sameAs`. A machine-readable identity map.
- 3.Ensure consistency (NAP and description). Identical name, address, contact details and description across all profiles (Google Business Profile, LinkedIn, industry directories). Mismatches weaken Google's "confidence".
- 4.Get into Wikidata. A direct Knowledge Graph source with softer criteria than Wikipedia - I describe how to create a correct record step by step in the post on Wikipedia, Wikidata and the Knowledge Panel.
- 5.Earn external corroboration. Mentions in industry media, quality profiles, talks - the more consistent, independent sources, the stronger the entity.
- 6.Build co-occurrence. Consistently tie the brand to its topic in content (you + your topics, you + your technologies) so the model cements the associations.
Wikidata step by step
Wikidata is the most underrated Entity SEO lever, because it feeds the Knowledge Graph directly and has a lower bar than Wikipedia. How to approach it:
- Check notability. Wikidata requires the entity to be "clearly identifiable" and described by a reliable, independent source - not self-promotion. If you have media mentions, talks, publications - you have material for references.
- Add solid references. Back every meaningful claim (what the company is, when it started, what it does) with an external source. Unsourced items get removed.
- Link back. From the Wikidata item, point to the official site, and from the site - via `sameAs` - to the Wikidata item. That closes the corroboration loop.
- Don't spam. Wikidata is a moderated community. Artificial, self-promotional items get reverted - build it properly.
Entity home - how to write a page a machine understands
The "About" page isn't corporate fluff - it's a technical identity document. A strong entity home:
- Answers directly: who you are, what you do, for whom, where you operate, why you can be trusted.
- Includes specifics and data - years of experience, number of projects, named technologies and specializations (this builds co-occurrence).
- Is marked up with schema (`Organization`/`Person` + `ProfilePage`) and ties the author via `@id`.
- Links to corroboration - profiles, publications, talks, Wikidata via `sameAs`.
It's from this page that models and Google "learn" your entity - so it must be unambiguous, fact-rich and consistent with the rest of the web.
How to measure entity strength
An entity isn't a single-number metric, but several signals show progress:
- Brand SERP - type your brand name into Google. Is there a knowledge panel? Are the descriptions and links correct? It's the fastest thermometer of Google's "confidence".
- Knowledge panel - its appearance is proof the entity is in the graph and sufficiently corroborated.
- Autocomplete and "People also ask" - does Google associate the brand with the right topics.
- An AI test - ask ChatGPT/Perplexity "what does [brand] do" and check whether the answer is accurate (this is also part of an AI visibility audit).
The most common Entity SEO mistakes
- Starting with Wikidata instead of your own site - external signals have nothing to "stick" to without a consistent entity home and schema.
- Inconsistent data (different names, addresses, descriptions) blurring the entity.
- A thin "About" page with no specifics - the model has nothing to build understanding from.
- Treating schema as a "boost" - it's a clarity foundation, not a ranking factor.
- Self-promotional, unsourced Wikidata items - reverted by the community.
Important caveat: schema is NOT a ranking factor
This is the most misstated point. Google (John Mueller, Danny Sullivan) repeatedly confirms: structured data doesn't directly raise rankings, and there is no special schema required for AI Overviews or AI Mode - a page simply has to be indexed and eligible for results. Schema's role is to help understand the entity and grant rich-result eligibility, not a "boost". Likewise, "reciprocal `sameAs`" or "building an entity graph increases AI citations" is sensible identity-organizing practice, but not a provider-confirmed mechanism - treat it as a clarity foundation, not a promise.
The entity and the Knowledge Graph in practice (knowledge panel)
A knowledge panel is generated automatically when the entity is in the graph and has enough corroborated information on the web - you can't buy or "order" one. Once it exists, you can claim it after verification and suggest edits. The most effective path is building consistent, corroborated facts about the entity (including via Wikidata) to raise Google's "confidence".
Entity SEO is also the bridge between SEO and GEO: AI models reason in entities, so a consistent, well-described entity increases the chance AI "knows who you are" and cites you.
---
I design brand entity architecture as part of knowledge graph engineering and Entity SEO and full AI-GEO. I teach it in the SEO & GEO course. Get in touch - I'll start by auditing how your entity looks to Google today.
Worth reading next:
/// RELATED_SERVICES
Need these concepts implemented? Explore the services related to this topic.
/// SOURCES
/// RELATED_RECORDS
Brand Reputation in AI - What to Do When ChatGPT Says Something Untrue About Your Company
In a July 2026 Searchable study, 93% of 1,704 brands had at least one fact wrong or missing in AI answers, and small businesses got fabrications far more often. Where models get untrue facts about a brand, how to audit them, five correction paths and what the law says when an error costs a contract.
Image SEO in the AI Era - Google Lens, Visual Search and Graphics Models Cite
Google Lens now processes close to 20 billion visual searches a month, and Circle to Search runs on over 580 million Android devices - images stopped being an afterthought to text and became a search surface of their own, with its own rules. Here's what alt text actually does (and doesn't do), how to wire images into ImageObject structured data, why an infographic built as HTML beats an infographic-as-image for citability, which file format to pick in 2026, and what the EU AI Act says about labeling AI-generated graphics.
Bing, Microsoft Copilot and IndexNow - the Forgotten Ecosystem Feeding AI Answers
On July 1, 2026, Fabrice Canel, Bing's longtime head of crawling and indexing and the chief architect of the IndexNow protocol, retired from Microsoft. It's a good moment to notice how much of today's AI-generated answers actually run on the infrastructure he built. The whole SEO world is debating Google AI Overviews, while Microsoft Copilot entirely, and ChatGPT substantially, pull sources from the Bing index - and since February 2026, Bing itself shows you when you're being cited. Here's how the ecosystem works, how to configure it, and how much traffic it actually delivers.
Signal received?
Terminate
Silence
Initiate protocol. Establish connection. Let's build something loud.