Google Knowledge Graph:
What It Is and How Panels Work

The Google Knowledge Graph is Google’s database of facts about people, places, and things. Google Search uses it to answer factual questions and show knowledge panels about entities. General knowledge panels are generated automatically; Google does not manually create them on request.

What is the Google Knowledge Graph?#

Google introduced the Knowledge Graph in May 2012 to help Search understand real-world subjects and the relationships between them. An entity is an identifiable subject, and Google’s developer documentation describes entities as the graph’s nodes, with types including people, places, organizations, books, movies, and events.

Facts describe those subjects, while relationships connect them; Google’s launch explanation describes using this information to distinguish subjects that share a name, summarize relevant facts, and help people discover related subjects.

Where does the information come from?#

Google’s explanation of its data sources identifies material shared across the web, open-source databases, and licensed databases. Wikipedia is one commonly used source among many, and Google also licenses information for subjects such as music, sports, and television.

Google’s Knowledge Graph help page also lists factual information received from content owners, including suggested changes to panels they have claimed. For knowledge panels specifically, Google combines web information with authoritative data partners and verified feedback, and updates panels automatically as information on the web changes.

A panel’s contents vary by subject, but Google’s description of typical panel fields includes a title, description, images, key facts, and links to official websites or social profiles.

Knowledge Graph, knowledge panel, structured data, and Business Profile#

The distinction matters for local information: Google’s explanation of Business Profile management identifies opening hours and contact phone numbers as details local businesses can manage through that separate process.

Can you create a Google knowledge panel?#

Google’s documented creation process is automatic: its Search systems generate knowledge panels when sufficient information is available on the open web. Its panel feedback policy says Google does not manually create or delete general knowledge panels.

Claiming starts with finding an existing panel and verifying an association with the entity, establishing who can represent the subject and submit feedback.

A panel can also stop appearing: Google says its automated systems decide when, where, and how to display one based on relevance to the query, so an earlier appearance does not establish a permanent display for that search.

How to make organization information clearer to Google#

Google says Organization structured data can help it understand administrative details and distinguish organizations with similar identities, while some properties can influence visual elements, including the logo shown in Search and a knowledge panel.

Start with visible information on the website: Google’s structured data introduction says markup must describe the page where it appears, and warns against empty pages created solely to hold markup or information hidden from readers. Publish the identity information that the markup will describe, and maintain the page and markup together.

For an organization, Google’s implementation guidance recommends the homepage or one page describing the organization, such as an About page, and the most specific applicable Organization subtype. Its property definitions cover:

• name and a genuine alternateName;
• legalName when the registered name differs from the public name;
• url for the organization’s website;
• sameAs for relevant profiles on other websites;
• logo for a representative image; and
• applicable address and contact details.

There are no required Organization properties; Google recommends including relevant information that applies to the organization. Keep it current and representative of the visible page, as required by Google’s structured data quality guidelines.

Validate the markup with the Rich Results Test described in Google’s introduction, then check technical access with the URL Inspection tool recommended in Google’s guidelines. Those guidelines require access for Googlebot and explain that correctly implemented markup does not guarantee a special search appearance.

How to claim and correct an existing panel#

The subject or an official representative can follow Google’s verification instructions: find the panel in Search, select the claim option, and verify through an associated official website or profile. Google notes that some panels cannot currently be claimed; their Feedback option remains available for suggested corrections.

For verified corrections, Google’s submission instructions require the verified Google Account, the Suggest edits option, an explanation of the error, and publicly accessible supporting URLs. Submit each item separately; Google checks suggestions against public information and may decline changes it cannot confirm.

Two fields have specific limits:

• Description: Google directs corrections to the original information source; with strong evidence, it may remove an inaccurate description, but cannot write a custom replacement.
• Featured image: A replacement can be suggested when an image already exists, but Google cannot add an absent image on request.

People who are not official representatives can also submit corrections through a panel’s Feedback link.

How to search the Knowledge Graph through an API#

Google’s legacy Knowledge Graph Search API overview describes entity lookup, search completion, and content annotation. It returns matching entity records rather than a graph of interconnected entities, and its read-only design provides no route to add or edit a public entity.

The legacy entities.search reference documents searches by query or entity ID, with filters for language and entity type. Its JSON-LD responses can include an entity’s identifier, name, type, description, image, and official website URL.

The reference defines resultScore as how well an entity matches the request constraints; it does not define the score as a threshold for receiving a knowledge panel.

For new integrations, Google’s legacy API documentation directs developers to Cloud Enterprise Knowledge Graph and warns against making the legacy service a production-critical dependency. The current Google Cloud search documentation labels the service as Preview, recommends the Advanced edition for new projects, and documents search and lookup operations. It also states that the Cloud search API returns individual matching entities rather than interconnected graphs.