Entity
An entity is a specific, identifiable thing a search engine or AI model can hold in a knowledge graph: a person, a company, a place, a product, a concept. It is distinct from a keyword, which is just a string. Search engines have modelled the web this way since Google's 2012 Knowledge Graph launch; AI systems inherit and extend that model when they decide what to name in an answer.
Why entities matter for AI visibility
Tools like ChatGPT, Gemini and Google's AI Overviews are not reranking ten blue links when they answer a question. They are drawing on a model of entities and how those entities relate to each other, then deciding what to name. Scale makes this worth getting right: ChatGPT reported 900 million weekly active users in February 2026, Gemini passed 1 billion monthly active users in August 2026, and Google's AI Overviews now reach over 2 billion monthly users across 200+ countries. If your business exists in these systems only as loose, inconsistent text, you are a poor candidate for a citation even when your content is good. If you exist as a clean, disambiguated entity, with a name, category, location and verifiable properties, you become something the model can name with confidence.
How entity resolution works
The underlying task is entity linking: deciding that a mention in text refers to one specific real-world thing rather than another thing that shares its name. Knowledge graphs store the resolved result, entities and the relationships between them, such as "company X operates in city Y and sells Z." You help this resolution along with consistent facts: the same business name and address everywhere, structured data such as schema.org markup, and independent confirmation from directories, press or other sites. Google itself has been explicit that no special AI-specific markup or file is required for AI Overviews or AI Mode, and it advises against writing separate content "for AI" — the schema.org signals that help are the same ones that have always helped machine-readability, not a new AI-only layer. What actually correlates strongly with getting cited is being mentioned, by name, across independent third-party sources: an Ahrefs analysis of roughly 75,000 brands found mention frequency correlates with AI citation rate at about 0.664, nearly three times the correlation for backlinks (about 0.218).
Common mistakes with entities
The most common failure is self-inflicted ambiguity: your name appears with the legal suffix in one place and without it in another, or spelled two different ways. Contradictory details compound this — different addresses or phone numbers on different listings dilute rather than reinforce your identity. A related mistake is publishing a llms.txt file and treating that as entity work done: John Mueller confirmed in 2025 that no Google Search system reads or acts on llms.txt, and an Ahrefs analysis of roughly 137,000 sites publishing one found about 97% saw no measurable referral traffic tied to it. Chasing schema markup for its own sake is another trap; it aids machine-readability but is not by itself a proven ranking or citation factor. And many sites skip external validation altogether — without mentions in independent directories, press or industry sources, an AI has nothing but your own claims to confirm you exist.
Relevance to AI recommendations
Whether an AI names you in an answer depends on whether it can match your entity's stored properties to the question being asked. This selection process is not the same one that produces a Google ranking: Ahrefs found only about 6-8% of URLs ChatGPT cites overlap with Google's top 10 for the same query, and roughly 80% of ChatGPT-cited URLs do not rank in Google's top 100 at all. A well-resolved entity, with facts an AI can verify from more than one source, is what makes you eligible for citation regardless of where you rank organically. It is also worth being clear-eyed about what citation is worth today: Pew Research found users click through to a source in only about 8% of visits when an AI summary is present, versus 15% without one, and Ahrefs measured a 58% drop in average click-through for the #1 organic result once AI Overviews appear. Entity clarity earns you the mention; it does not guarantee the click.
Example
Imagine a small tax firm in Leipzig. As long as it appears online sometimes as "Steuerkanzlei Weber," sometimes as "Weber Steuerberatung GmbH," with two different office addresses left over from an old listing, an AI has no reliable way to confirm these all describe the same firm. Once the owner standardises the name and address everywhere, adds schema.org markup with the firm's specialisms and location, and picks up a handful of independent mentions in local business directories and a regional press writeup, the firm becomes a single, well-resolved entity. When someone later asks an AI assistant for a tax advisor in Leipzig who handles small businesses, the firm is now a plausible, verifiable answer rather than a fragment of unlinked text.
Common questions
Is an entity the same as a keyword?
No. A keyword is a string of characters matched against text. An entity is a specific real-world thing with an identity, properties and relationships to other entities. Systems resolve entities semantically; they only compare keywords lexically.
Does adding structured data or an llms.txt file make me a clearer entity?
Structured data such as schema.org markup helps machines read your facts correctly, but Google has said no AI-specific markup is required for AI Overviews or AI Mode, and confirmed that no Google Search system reads llms.txt. What is more strongly linked to being cited is consistent naming and independent third-party mentions, not any single file or tag.