A uniquely identifiable thing — a person, brand, place, or concept — that a search engine can pin to one distinct record in its knowledge graph.
Last verified · 2026-09-03 · by Moe Ameen
In SEO, an entity is any thing or concept that is singular, unique, well-defined, and distinguishable — a person, a company, a product, a place, an event, an idea, even a color. What makes something an entity is not its category but its resolvability: a search engine can decide, confidently, that a given mention refers to this one specific thing and not to something else that happens to share the name. Entities are the atomic units of meaning in modern search — the recognized things that a keyword string only points at.
The concept exists because search moved from matching strings to understanding things. For most of SEO's history an engine inferred relevance from text patterns and links; it did not really know what a page was about. Google's Knowledge Graph, launched in May 2012 under the slogan "Things, not strings," changed that by holding real-world entities and the relationships between them, so search could reason about a thing rather than just return pages containing a phrase. Entity-based SEO is the practice that follows: optimizing to be understood as a distinct, recognized thing, and to be correctly related to the entities around it, rather than only to match an exact query.
By 2026 this is no longer a fine-grained distinction. The AI systems layered on search — AI Overviews, AI Mode, chatbot answers — reason over the same knowledge graph, so whether your brand exists as a recognized entity, how it is described, and what it is connected to directly shape whether an answer engine will name you. A page with strong keyword optimization but weak entity signals routinely loses a citation to a page from a brand the engine can actually resolve and trust, because an LLM cannot verify an entity it has no representation for.
Entity thinking entered mainstream search with two 2012–2013 releases. Google publicly launched the Knowledge Graph on May 16, 2012, describing it as a way to search for things, not strings — a database of real-world entities and their relationships that could power richer, more direct answers. The 2013 Hummingbird update, announced on September 26, 2013 on Google Search's fifteenth anniversary, rebuilt the core ranking algorithm to interpret the intent and context of a whole query rather than weighting isolated keywords, which is what let the Knowledge Graph's entity understanding feed everyday results.
The idea has intellectual roots older than Google's product — semantic-web and linked-data work had modeled the world as entities and relationships for years, and schema.org (2011) gave publishers a shared vocabulary to declare them. What changed over the following decade was scale and stakes: Google has described its Knowledge Graph as holding hundreds of billions of facts about billions of entities, and the arrival of AI answer surfaces trained on that graph turned entity recognition from a knowledge-panel curiosity into the prerequisite for being visible in AI search at all.
| Platform | Behavior |
|---|---|
| Google Knowledge Graph | The canonical entity database behind Search. Confident entities get a stable machine identifier (a MID) that is the same across languages and spellings, and can surface a knowledge panel. Google's AI Overviews and Gemini answers reason over this graph, so a brand's representation here shapes AI-search visibility directly. |
| Schema.org / structured data | The vocabulary you use to declare your own entities in JSON-LD. Organization and Person types with a sameAs array explicitly link your entity to its other profiles, handing engines a ready-made corroboration map. Schema does not rank you, but it makes your entity machine-readable and verifiable. |
| Wikidata / Wikipedia | Open, structured knowledge bases that engines lean on heavily to confirm an entity exists and what it is. An entry here is one of the strongest corroboration signals available, but it is earned through genuine notability, not declared — you cannot simply create one for a brand that has done nothing notable. |
| Bing / Copilot | Maintains its own entity graph feeding Copilot answers, and rewards structured, well-marked-up content when resolving and citing entities. The signals overlap heavily with Google's — consistent naming, corroborating profiles, structured data — even though the underlying graph is separate. |
| LLM answer engines (ChatGPT, Perplexity) | Rank on content quality, specificity, and authority rather than a published entity graph, but still favor sources they can resolve to a recognized, trusted entity. A weak or ambiguous brand entity is a frequent reason a well-written page fails to get named in a generated answer. |
The most useful reframe of entity SEO is that it is a disambiguation problem before it is a marketing problem. Search does not fail to cite most brands because they lack keywords — it fails because it cannot confidently decide who they are. So the work is unglamorous: pin down one canonical description of the thing and repeat it, identically, everywhere it appears, until an engine has no room left to doubt that all these mentions resolve to one record.
The trap in the AI era is that scaling content with generic tools actively erodes this. If every post drifts in voice and describes the brand a little differently — because a blank prompt returns the model's median house style, not yours — you are feeding engines an inconsistent, blurry entity at high volume, which is worse than publishing less. The point of forcing your specific identity into generation, the way a governed engine like Kompozy does through its [Persona Brief](/glossary/persona-brief) and a consistent persona identity, is not aesthetic polish. It is that the brand gets characterized the same way on every surface, so the corroborating signal an engine collects across your output stays coherent instead of dissolving into noise. Consistency is the entity signal; anything that quietly varies your description at scale is working against you.
An entity is any uniquely identifiable, distinguishable thing — a person, brand, place, product, event, or concept — that a search engine can pin to one unambiguous record in its knowledge graph. Entity-based SEO treats content as being about these recognized things rather than about keyword strings, because Google and the AI systems built on it reason in entities, not exact-match phrases.
A keyword is a string of characters a searcher types; an entity is the real-world thing that string points at, held with its attributes and relationships. "Apple" is a string that could mean the fruit, the label, or the company — the entity is the specific one an engine resolves from context. Keyword SEO matches the phrase; entity SEO makes you the recognized thing and connects you correctly to related things.
AI Overviews, AI Mode, and chatbot answers reason over the same knowledge graph that defines entities, so how your brand is represented as an entity shapes whether an AI answer names you. LLMs cannot verify a brand they have no representation for, so a page with strong keywords but weak entity signals often loses a citation to a page from an entity the engine can resolve and trust.
Give your brand one canonical home page that clearly states what it is; name and describe it identically everywhere it appears; corroborate that identity with consistent profiles and a sameAs array in your Organization schema; and publish substantive content covering your topic and the entities around it. Recognition is earned by corroboration — the same clear description of the same thing in enough trusted places for an engine to resolve it confidently.
No — it sits beneath it. You still target real queries, but you win them by being a recognized entity with clear topical coverage rather than by repeating a phrase. As search shifted from matching strings to understanding things, entity signals became the foundation that keyword relevance now rests on, especially for surfacing in AI-generated answers.