How to build your crypto brand's entity in the knowledge graph

What an entity actually is

An entity is a uniquely identified concept in the knowledge graph. Your brand has a name, a description, a category, properties, and relationships to other entities. The knowledge graph stores all of this as a node, not as a webpage.

The practical implication: search engines and AI engines route queries to entities, not URLs. When someone asks 'who is <YourBrand>,' the engine looks up the entity, then surfaces information about it. If the entity exists and is well-defined, you get a clean answer. If the entity doesn't exist, the engine falls back to whatever URL contains the brand name most prominently — which is often a competitor's comparison page.

For crypto brands, entity definition matters more than mainstream SEO because crypto brands often share names with unrelated companies (SUI the network vs SUI the cosmetics brand, BNB the token vs BNB the music genre tag).

The 4 sources AI engines trust most

From our entity-classification testing, AI engines reliably trust 4 sources when defining a crypto brand entity.

1. Wikidata. The structured-data backbone of Wikipedia and the knowledge graph. A Wikidata entry with proper P31 (instance of), P127 (owned by), P856 (official website) properties anchors your entity definitively.

2. CoinGecko and CoinMarketCap. Tied for second. AI engines treat these as crypto-specific authority sources, especially for tokens and exchanges. A claimed and verified profile on both is non-negotiable for any token issuer.

3. The brand's own schema. Your Organization or FinancialService schema with complete sameAs links. This is how the engine confirms 'the brand on this site is the same entity as the one on Wikidata and CoinGecko.'

4. Crunchbase and LinkedIn Company. Less crypto-specific but heavily trusted for entity properties like founders, founding date, headquarters, and funding history.

Wikidata entry setup

Wikidata is editable by anyone, but quality submissions stick and low-quality submissions get deleted. The acceptance criteria: notability (must be referenced in at least 2 independent reliable sources), neutral point of view, no promotional language.

For a crypto brand, the minimum viable Wikidata entry has these properties:

• Label: brand name in English
• Description: 10-15 word neutral description (e.g. 'decentralized lending protocol')
• P31 (instance of): match the right entity class (e.g. 'cryptocurrency', 'DeFi protocol', 'cryptocurrency exchange')
• P856 (official website): your canonical URL
• P3417 (Twitter username) if you have one
• P2002 (GitHub username) if applicable
• P571 (inception): founding date
• P127 (owned by) or P112 (founded by) if relevant
• At least 2 source references (P854, reference URLs)

Submission process: create an account at wikidata.org, search for your entity (might already exist), create or edit the item. Include a reference URL for every property claim. Property changes take effect immediately but indexing into the knowledge graph can take 1 to 4 weeks.

schema.org sameAs network

The sameAs array is how you tell AI engines 'this entity on my page is the same as the entity on these other authoritative sources.' Include at minimum these URLs in your Organization schema:

"sameAs": [
  "https://www.wikidata.org/wiki/Q123456",
  "https://www.coingecko.com/en/exchanges/yourbrand",
  "https://coinmarketcap.com/exchanges/yourbrand/",
  "https://www.crunchbase.com/organization/yourbrand",
  "https://www.linkedin.com/company/yourbrand",
  "https://twitter.com/yourbrand",
  "https://github.com/yourbrand"
]

The Wikidata URL is the most powerful entry. It's the canonical identifier in the knowledge graph. CoinGecko and CMC follow. The social URLs add coverage but each contributes less individually.

Verify every URL returns 200. Dead sameAs links are worse than no sameAs because they signal stale or auto-generated schema, which AI engines de-weight.

Handling name collisions

If your brand name collides with another entity (different industry, different language, etc), the entity disambiguation falls on your schema and Wikidata description.

Step 1: Make your description as specific as possible. 'Cryptocurrency exchange founded 2023, headquartered in Singapore' beats 'crypto trading platform.' Specificity is what tells the engine your entity is different from the entity with the same name.

Step 2: On Wikidata, use the disambiguation properties. P31 (instance of) should clearly distinguish you from the other entity. If you're a crypto wallet and another brand with your name is a clothing line, your P31 ('cryptocurrency wallet') versus theirs ('clothing brand') resolves the conflict.

Step 3: Include a disambiguation note in your homepage's Organization schema description. 'YourBrand is a cryptocurrency wallet, not affiliated with the clothing brand of the same name.' AI engines pick up these explicit disambiguations and route correctly.

Common questions