Google Knowledge Panel: What It Is, Types & How It's Generated

A knowledge panel is the information box Google Search displays when it determines that a query refers to a specific entity it holds in the Google Knowledge Graph. It normally appears at the right of desktop results or above results on mobile, and it summarizes the entity with a name, a type label, a short description, an image, and a set of facts and links appropriate to what kind of thing the entity is. Google generates panels automatically. There is no application process, no paid placement, and no guarantee that any given entity will get one.

What causes a panel to appear

Two conditions have to hold. Google must hold an entity that the query resolves to, and Search must judge the query to be about that entity rather than about documents that mention it. A search for a common two-word phrase that happens to be a company name will often return no panel, because the resolution is ambiguous. A search for the same company plus its city frequently does, because the additional term disambiguates.

This is why panels are less stable than people expect. The same query can produce a panel for one user and not another, panels can appear and disappear across weeks, and a brand can hold a graph entity with no panel to show for it. The entity and the panel are separate: the graph is data, the panel is a rendering decision made per query.

Panel types and where each gets its data

Panels are not one template. The type of the entity determines both the fields shown and, more importantly, which data sources feed it.

Panel type	Typical fields	Primary data path
Organization or brand	Description, logo, founded, founders, headquarters, key people, social profiles, stock ticker	Knowledge Graph, drawn from public sources, licensed data and the web
Person	Description, birth date, education, spouse, notable works, social profiles	Knowledge Graph, heavily reliant on public reference sources
Local business	Address, hours, phone, directions, reviews, photos, popular times	Google Business Profile, which the owner manages directly
Film, TV series, book, album	Release date, director or author, cast, ratings, where to watch or buy	Licensed domain databases plus the Knowledge Graph
Place, species, medical condition and similar	Domain-specific fact sets	Licensed and curated sources for the domain

The local business case is the one that behaves differently from the rest, and confusing it with the others causes most of the frustration in this area. A local business panel is largely a rendering of a Google Business Profile. The owner edits the underlying record directly, changes appear quickly, and the workflow is a product with a dashboard. An organization or person panel has no dashboard. Its content comes from Google’s graph, and the only inputs available are verification and suggestion, described below.

Knowledge sources

Google documents its panel sources in categories rather than as a weighted list. What it has stated publicly covers:

Publicly available sources on the open web, with Wikipedia named explicitly and Wikidata among the structured sources.
Data licensed from providers in specific domains, such as sports statistics, film and television metadata, and music catalogs.
Business information submitted through Google Business Profile.
Information supplied by verified representatives of the entity through the claiming process.
Facts extracted from web pages, including structured data those pages publish.

Google also states that it does not accept payment to create or change a panel, and that changes take time to appear because the underlying graph updates on its own schedule.

What Google does not document is any threshold or weighting. There is no published rule of the form “N independent sources produce an entity” or “a Wikipedia article produces a panel”. Statements of that kind in SEO writing are inference from observed cases.

The role of Wikipedia and Wikidata

Practitioners notice that entities with Wikipedia articles very often have panels, and conclude that Wikipedia is the switch. That reading is too strong in one direction and too weak in another.

Too strong, because plenty of entities hold panels with no Wikipedia article at all. Local businesses, many private companies, authors with catalog records, and musicians with distributor metadata all appear without one. Wikipedia is a source Google names, not a gate it documents.

Too weak, because Wikipedia and Wikidata do something valuable that is easy to misdescribe. They are highly structured, heavily cross-referenced, and permissively licensed, which makes them efficient corroboration for facts Google can also find elsewhere. A Wikipedia article is itself built from independent published sources, so an entity that qualifies for one has, by construction, the third-party coverage that makes it identifiable. The article and the panel share a cause.

The accurate framing, and it should be labeled as practitioner inference rather than documented behavior, is that Wikipedia and Wikidata serve as corroboration rather than as an intake form. Creating a Wikidata item for a subject with no independent coverage does not manufacture an entity; it produces an item likely to be deleted under Wikidata’s notability guidelines, which require either a serious and publicly available reference, a structural role in the data, or a clearly identifiable subject.

Claiming a knowledge panel

For entity types other than local businesses, Google offers a verified representative flow, surfaced as a “Claim this knowledge panel” link on the panel itself. The sequence Google documents:

Sign in to a Google account and find the panel in Search.
Select the claim link on the panel.
Verify an official presence associated with the entity. Google’s documentation lists routes including a website verified in Search Console, an official YouTube channel, and official profiles on supported social platforms.
Once verification is complete, the account is marked as a verified representative of that entity.

Verification does not confer editing rights. A verified representative can suggest changes, including a featured image, and Google states that suggestions from verified representatives are reviewed, without committing to accepting them. Google decides what the panel says. The value of verification is a faster and more legible route for corrections, not control.

Two constraints matter before starting. The claim link only appears if Google already holds and displays the entity, so claiming is a step for entities that have panels rather than a way to obtain one. And verification requires an official presence Google can check, which is one more reason to keep the entity’s website, YouTube channel and social profiles consistently named and interlinked.

Suggesting an edit

Anyone signed in can use the feedback link on a panel to report an error. This is the route for a wrong founding date, a wrong image, a description that names the wrong company, or a merged entity that has conflated two similarly named things. Suggestions perform better when the correction is factual, specific, and supported by a source Google can verify independently of the entity itself. Reports asking for a description to be reworded for marketing reasons generally go nowhere, because a panel description summarizes what sources say rather than what an entity would prefer them to say.

Conflation between two entities with the same name is the most common serious defect and the hardest to fix, because it is a resolution error in the graph rather than a wrong field value. The lever that works over time is differentiation in the sources: distinct names where possible, distinct locations and identifiers stated in markup, and sameAs links pointing at separate external records.

What a panel is not

A knowledge panel is often confused with three other Search features that look similar and work differently.

A featured snippet is an extracted passage from a ranking web page, shown at the top of results with a link to its source. It is document-derived, it changes with the ranking page, and it carries no entity association.

The local pack is a map plus a set of nearby business listings for a query with local intent. It draws on Google Business Profile data, as local panels do, but it lists multiple businesses rather than describing one entity.

A site name and favicon in a result, or the newer merchant and organization result treatments Google supports through markup, are result decorations, not panels. Markup can influence those. Markup cannot conjure a panel.

For developers

Panels are not available through any API. The Knowledge Graph Search API queries a subset of entities and returns a fixed set of fields, and its results neither predict nor reflect panel presence. Anyone building an entity monitoring system therefore has to reconcile two separate signals: an API lookup that shows whether Google holds a matching entity at all, and SERP observation that shows whether a panel renders for a given query, location and device.

For entity reconciliation work in a pipeline, the open bases are the practical source. Wikidata exposes the full graph over SPARQL, so an entity’s identifiers, types and relationships can be pulled directly, and identifiers such as VIAF, ISNI, ORCID and official register numbers can be used as join keys. The modeling underneath is described at what is an entity and triples: subject, predicate, object.

For SEOs

The work that plausibly matters divides into three groups, in decreasing order of documentation.

Documented. Publish valid structured data describing the entity, using JSON-LD with an Organization or Person node, a stable @id, and sameAs links to external records that describe the same entity. Google documents that it reads structured data and lists it among its sources.

Consistent with what Google documents, though the effect is unquantified. Keep one canonical entity home page. Use one name form everywhere. Make sure the description a site gives itself matches the description independent sources give it, because a mismatch gives Google nothing to corroborate.

Practitioner inference. That third-party coverage volume and quality correlate with entity representation. That Wikipedia and Wikidata presence tends to accompany panels. That conflated entities separate over time as sources differentiate. These are observed patterns worth acting on, and they are not published mechanisms, so they should not be sold as guarantees.

The step-by-step version of this is at how to get a knowledge panel, and the wider program at entity SEO.

FAQ

What is a Google knowledge panel?

A Google knowledge panel is the information box Search shows when it determines a query refers to a specific entity in its Knowledge Graph. It appears at the right of desktop results or above mobile results and summarizes the entity with a name, type, description, image and relevant facts. Google generates panels automatically and does not accept payment for them.

How do I claim my knowledge panel?

Sign in to a Google account, find the panel in Search, and select the “Claim this knowledge panel” link. Google then asks you to verify an official presence for the entity, through routes including a website verified in Search Console, an official YouTube channel, or official social profiles. Verification lets you suggest changes; it does not let you edit the panel directly.

Do I need a Wikipedia page to get a knowledge panel?

No. Many entities with panels have no Wikipedia article, including local businesses, private companies and musicians with catalog metadata. Wikipedia is a source Google names, not a documented requirement. An article helps mainly because qualifying for one means the entity already has the independent published coverage that makes it identifiable.

Why did my knowledge panel disappear?

Panels are rendering decisions made per query, not permanent properties of an entity, so they can appear and disappear as Google’s confidence in the query-to-entity match changes, as the underlying graph data updates, or as source coverage shifts. Google does not notify entity representatives about panel changes and does not document the thresholds involved.

Sources and further reading

Google. Knowledge panel help center (claiming, verification and suggesting changes). https://support.google.com/knowledgepanel
Google. A reintroduction to Google’s Knowledge Graph (May 2020). https://blog.google/products/search/about-knowledge-graph-and-knowledge-panels/
Google Business Profile Help. https://support.google.com/business
Wikidata. Notability. https://www.wikidata.org/wiki/Wikidata:Notability
Google Search Central. Introduction to structured data markup. https://developers.google.com/search/docs/appearance/structured-data/intro-structured-data

Google Knowledge Panel: What It Is, Types & How It’s Generated