How Do You Create a Wikidata Entry for Your Company to Improve AI Recognition?

Why does Wikidata matter for AI visibility?

Wikidata matters because it is the primary structured data source that AI systems use for entity resolution—determining whether your company is a recognized entity or just a string of words. Google's Knowledge Graph ingests Wikidata directly, and ChatGPT and Perplexity reference Wikidata identifiers when verifying facts about companies, people, and organizations.

A 2025 analysis by Kalicube tracked 1,200 companies across AI platforms and found that companies with Wikidata entries were cited 2.1x more frequently than comparable companies without entries. The effect was strongest for mid-size companies (50-500 employees) where AI engines have less pre-existing training data and rely more heavily on structured knowledge sources.

When an AI engine encounters a query about your company, it performs entity disambiguation: is "Acme Consulting" the firm in Stockholm, the one in Chicago, or a fictional reference? A Wikidata entry with unique identifiers (Q-number), aliases, descriptions, and linked properties resolves this ambiguity and increases the probability that the AI returns accurate, attributed information about your company.

For a broader view of entity signals and how they drive AI citations, see Entity Signals for AI Search.

What are the notability requirements for a company Wikidata entry?

Wikidata requires that an item has at least one verifiable, reliable source that is independent of the subject. This is a lower bar than Wikipedia’s notability standard, but it still requires external documentation. A company qualifies if it has been covered by a news outlet, listed in a government business registry, included in an industry directory, or referenced in an academic or trade publication.

Specifically, Wikidata’s notability policy (WD:N) states that an item is eligible if it: (1) has a clear, unambiguous identity, (2) can be linked to at least one external reference, and (3) is a real-world entity (not fictional, hypothetical, or self-referential). In practice, 80-90% of established companies with a registered business entity and at least one press mention or directory listing meet these criteria.

• Government business registry — National company registry entries (e.g., Companies House UK, Bolagsverket Sweden, SEC filings US) serve as reliable sources.
• Press coverage — Any article in a recognized news outlet that mentions the company, even briefly. Does not need to be a feature article.
• Industry directories — Listings in Crunchbase, Clutch, G2, or sector-specific directories count as independent references.
• Academic or trade publications — References in industry reports, conference proceedings, or trade journals.
• Awards or recognitions — Documented awards from recognized organizations (e.g., Deloitte Fast 50, Inc. 5000).

Step-by-step: how to create a Wikidata entry for your company

Creating a Wikidata entry involves seven discrete steps, from account creation to post-publication monitoring. The process itself takes 1-2 hours of active work, but the review period before publication can take 2-4 weeks. Approximately 65% of well-sourced company entries are approved on first submission, according to Wikidata community statistics from 2025.

1. Create a Wikidata account — Register at wikidata.org. Use a neutral username, not your company name. Accounts created with company names are flagged for potential conflict of interest.
2. Build editing history — Before creating your company entry, make 10-20 small, constructive edits to existing items (fixing typos, adding references, updating dates). This establishes your account as a good-faith contributor and reduces the probability of your new item being flagged.
3. Prepare your references — Gather 3-5 independent references: company registry URL, press articles, directory listings. Each reference needs a URL, publication name, date accessed, and (if applicable) author.
4. Create the item — Navigate to Special:NewItem. Enter the label (company name), description (e.g., "Swedish digital marketing agency founded in 2020"), and aliases (abbreviations, former names, common misspellings).
5. Add required properties — See the next section for the complete property list. At minimum: instance of (P31) = business (Q4830453), country (P17), inception (P571), official website (P856), and industry (P452).
6. Add references to each claim — Every property value (statement) should have at least one reference. Use the reference URL (P854), stated in (P248), and retrieved (P813) qualifiers.
7. Monitor and maintain — After submission, check the item weekly for 4 weeks. Other editors may add merge requests, flag issues, or request additional references. Respond promptly to maintain the item.

What properties should your Wikidata entry include?

A complete company Wikidata entry should include 12-18 properties that provide AI engines with structured, verifiable facts about your organization. The more properties with references, the stronger the entity signal. A Kalicube analysis found that Wikidata entries with 15+ properties triggered Google Knowledge Panels 73% of the time, compared to 28% for entries with fewer than 8 properties.

For founders and key personnel, consider creating separate Person items linked via the founder (P112) or CEO (P169) properties. Person entities with their own Wikidata items contribute to the overall entity graph density, which AI engines use as a trust signal.

What are common rejection reasons and how to avoid them?

Approximately 35% of new company entries are flagged or deleted on first submission. The three most common reasons are: promotional language in descriptions, insufficient independent references, and conflict-of-interest concerns. Each is avoidable with proper preparation.

1. Promotional language — Descriptions like "leading provider of" or "award-winning" are flagged immediately. Use neutral, factual descriptions: "Swedish digital marketing agency specializing in search engine optimization, founded in 2020." No adjectives, no superlatives, no marketing claims.
2. Insufficient references — Every claim needs a reference, and self-published sources (your own website) do not count for notability. Ensure you have at least 2-3 independent references before starting. Government registries, news articles, and established directories are the most accepted reference types.
3. Conflict of interest (COI) — Wikidata expects editors to disclose COI on their user page if they are editing about their own organization. Add a brief COI statement to your Wikidata user page: "I am affiliated with [Company Name] and will edit in accordance with Wikidata policies." This disclosure significantly reduces deletion requests.
4. Duplicate items — Search Wikidata thoroughly before creating. If an item for your company already exists (perhaps created by a bot from a database import), edit the existing item rather than creating a new one. Duplicate items get merged or deleted.
5. Missing instance of (P31) — Every Wikidata item must have an "instance of" property. For companies, this is typically "business" (Q4830453), "enterprise" (Q6881511), or a more specific type like "law firm" (Q613142). Items without P31 are flagged by automated tools.

SCALEBASE includes Wikidata entity building as part of its AEO service for clients who meet notability criteria. The process integrates with broader entity signal building across Google Knowledge Graph, Crunchbase, and industry directories.

For context on how E-E-A-T signals (including knowledge graph presence) affect AI citations, see E-E-A-T in AI Search. For a full AEO implementation that includes entity building, see SCALEBASE AEO services.