If you've never looked at a structured knowledge base entry directly, it can be a strange experience — there's no narrative, no marketing copy, nothing designed to be "read" in the conventional sense. That's because it isn't designed for human readers. It's designed for machines, and understanding its structure helps explain why it matters so much for AI visibility.
An identifier, not a name
Every entity in a structured knowledge base has a unique identifier — a code that refers to that specific entity, regardless of what it's called in any particular language or context. Names can change, be translated, or be ambiguous (many things share the same name); identifiers don't have that problem. This is part of why these records are so useful to machines — there's no ambiguity about which "thing" is being referenced.
A type
Every entity is categorised by type — is it a business, a person, a place, an organisation, a creative work, a product? This categorisation isn't just a label; it determines what kinds of properties and relationships are expected for that entity, and how other systems should interpret it.
Properties
Properties are the specific facts about the entity — when something was founded, where it's located, what category it belongs to, who operates it. Each property is itself often a reference to another structured value or entity, rather than free text — which is what allows these facts to be queried, compared, and reasoned over programmatically.
Relationships to other entities
Perhaps most importantly, entities are connected to other entities — a business is located in a particular place; that place is part of a particular region; that region is part of a particular country. These relationships form a web — quite literally a "graph" — that lets systems understand context: not just what something is, but how it fits into everything else.