Google's Knowledge Graph: How It Works and How to Check Your Brand

A question I hear a lot from business owners goes like this: “Why does Google show a nice box for my competitor, but nothing for us?” Many assume the box is something you pay for, or something a developer adds with a bit of code.

Neither is true. That box, the knowledge panel, is only the visible tip of something much bigger: Google’s Knowledge Graph. It is Google’s database of “things” (people, companies, places, products) and how they connect. If Google doesn’t clearly understand your brand as a “thing,” it has nothing to put in the box. And increasingly, it has less to work with when its AI features describe you.

Let me explain how the Knowledge Graph works in plain words, how to check whether your brand is in it using Google’s own API, how to read the results, and what to do if your entity is missing or wrong.

What Google’s Knowledge Graph actually is

Google introduced the Knowledge Graph in May 2012. In the launch post, Amit Singhal, then head of Google Search engineering, said it held about 500 million objects and more than 3.5 billion facts about them and the relationships between them. He described the goal with a phrase SEOs still use today:

“...understands real-world entities and their relationships to one another: things, not strings.”

— Amit Singhal, Google, Introducing the Knowledge Graph: things, not strings

“Strings” are words typed into a search box. “Things” are what those words refer to. When you search “Apple,” Google needs to know whether you mean the fruit or the company. The Knowledge Graph is how it keeps those apart.

Three terms will help you through the rest of this guide:

• Entity: a single “thing,” such as your company, your founder or your product.
• Attribute: a fact about an entity, such as founding date, headquarters or website.
• Relationship: a link between entities, such as “Jane Smith is the founder of Smith Dental.”

The graph has grown enormously since 2012. The Ahrefs guide to Google’s Knowledge Graph (updated May 2026) cites a Kalicube estimate of over 1.6 trillion facts about 54 billion entities, though no date is given for that figure. Ahrefs also reports, citing Search Engine Land, that Google removed over three billion entities in a single week in June 2025. I treat both as direction rather than exact counts, because Google itself only says the graph holds “billions of facts.” The cleanup is the more useful lesson: weak, poorly supported entities can disappear.

If you want the writing side of this topic, my guide to entity SEO covers how to describe your brand so search engines understand it.

Where Google gets the facts about your brand

Google’s help page How Google’s Knowledge Graph works is short but useful. It says:

“Facts in the Knowledge Graph come from a variety of sources that compile factual information.”

— Google, How Google’s Knowledge Graph works

The same page lists the types of sources: public sources across the web, data Google licenses (for things like sports scores and weather), and information from content owners, including people who claim a knowledge panel and suggest changes. In 2012, Google named Freebase, Wikipedia and the CIA World Factbook as early sources.

For a typical business, this is where Google most likely picks up facts about you:

Source	What it can tell Google	How much control you have
Your own website (About page, Organization schema)	Official name, logo, founders, social profiles, contact details	Full
Google Business Profile	Address, phone, hours, category, reviews	Full, once verified
Wikidata	Structured facts in a format machines read easily	Partial; you must follow its notability and editing rules
Wikipedia	Description and history	Low; you should not write your own article
News and trade press	Confirmation that you exist and what you are known for	Indirect, through digital PR
Directories and social profiles	Consistent name, address, links	High

The key word is consistency. When your site, your Business Profile, LinkedIn and trade directories all say the same thing, Google can be confident. When one says “Smith Dental Care LLC,” another “Smith Family Dentistry” and a third lists an old address, Google has to guess, or gives up. I cover Wikidata in detail in Wikidata for brands.

Knowledge Graph vs knowledge panel: not the same thing

People use these terms as if they were one thing. They aren’t.

• The Knowledge Graph is the database behind the scenes.
• A knowledge panel is one way Google displays some of that data, usually the box on the right side of desktop results.

So your brand can be in the Knowledge Graph without having a visible panel. Google may know you exist but not feel confident enough, or not see enough search interest, to show a box. The reverse is not possible: no entity, no panel.

This matters because the fix is different. If you have an entity but no panel, you work on confidence and demand. If you have no entity at all, you start with the basics. For claiming and editing a panel you already have, see my knowledge panel guide.

How to check your brand with the Knowledge Graph Search API

Here is the practical part. Google offers a free Knowledge Graph Search API that lets anyone look up entities. You don’t need to be a developer. You need a free API key and a web browser.

1. Create a Google Cloud project at console.cloud.google.com (free).
2. Enable the “Knowledge Graph Search API” in the API library.
3. Create an API key under Credentials. Restrict it to this API.
4. Paste this URL into your browser, replacing the brand name and key:
   https://kgsearch.googleapis.com/v1/entities:search?query=Your+Brand+Name&limit=10&indent=True&key=YOUR_API_KEY
5. Try variations: your legal name, your short name, and your name plus city.
6. Filter by type if you get noise, by adding &types=Organization or &types=Person for founders.

A few facts from Google’s API reference: the limit parameter defaults to 20 and can go up to 500; there is a prefix option that matches the start of names; and you can look up a known entity directly with the ids parameter. The usage limits page allows up to 100,000 read calls per day per project at no charge, which is far more than a brand check needs.

If you prefer not to touch an API key, the Ahrefs guide points to free third-party checkers that use the same API. They are handy, but I like the direct method because you see the raw answer.

One honest caveat. Google says plainly on the API overview page:

“This API is not suitable for use as a production-critical service.”

— Google for Developers, Knowledge Graph Search API

It is also read-only, and Google says it is moving this API to Cloud Enterprise Knowledge Graph. So use it as a diagnostic window, not a perfect copy of everything Google knows.

Reading the result: kgmid IDs, types and resultScore

The API answers in JSON (a structured text format). Here is a simplified, hypothetical result for a dental practice:

{
  "@type": "EntitySearchResult",
  "result": {
    "@id": "kg:/g/11example12",
    "name": "Smith Dental Care",
    "@type": ["Dentist", "Organization", "Place", "Thing"],
    "description": "Dental clinic in Austin, Texas",
    "url": "https://www.smithdentalcare.example/"
  },
  "resultScore": 412.7
}

Here is what each part tells you:

Field	What it means	What to look for
@id	The entity’s ID, often called a kgmid (machine ID)	Save it. It is your entity’s permanent reference. The documentation example looks like kg:/m/0dl567
name	The name Google uses	Should match your official brand name
@type	The schema.org types Google assigned	A dentist listed only as “Thing” means weak understanding
description	A short label	Should describe what you do, not something outdated
detailedDescription	A longer description, often from Wikipedia	Many small businesses won’t have one; that is normal
url	Official website	Should be your current domain
resultScore	How well the entity matched your search	Compare it with other results in the same search, not across searches

Google’s reference defines resultScore as “an indicator of how well the entity matched the request constraints.” It is not an authority score or a “brand strength” number, even though some tools present it that way. A score of 400 in one search and 40 in another don’t compare. Within one search, though, a large gap between your result and the next one is a good sign that Google clearly links your name to your entity.

Also note: according to Google Cloud’s migration guide, the newer Cloud version of this API removes the resultScore field entirely. So don’t build long-term reports around it.

Four results you might see, and what each means

• Your brand appears first, correct type and URL. You have a healthy entity. Focus on enriching it.
• Your brand appears, but with the wrong type or an old URL. Google knows you but has outdated facts. Go to the fixing section below.
• Another business with a similar name appears first. You have an ambiguity problem. Strengthen your distinct details (location, category, founders).
• Nothing relevant appears. Google may not see you as an entity yet, or the API simply doesn’t show you. Build the basics first.

How the Knowledge Graph connects to AI Overviews and AI Mode

This is where the topic gets more urgent. The Ahrefs guide says AI Overviews can use the Knowledge Graph to identify and verify the entities in a question and to ground answers in established facts. It links to a Google document about AI Overviews and AI Mode as support.

I want to be careful here. Google’s own developer page on AI features and your website (updated December 2025) does not mention the Knowledge Graph at all. It says there are no special requirements to appear in AI Overviews or AI Mode beyond being indexed and eligible for a snippet. Google’s generative AI optimization guide (updated July 2026) also says no special schema markup is needed.

So how do I put the two together? In brief: Google does not ask you to “optimize for the Knowledge Graph” to get into AI answers, but its systems clearly need to know who you are to describe you correctly. When someone asks AI Mode “Is Smith Dental Care good for kids?”, the system has to connect that name to the right business first, and then run the related searches I describe in my guide to query fan-out. A clean, consistent entity makes that easier. A messy one can lead to the AI mixing you up with someone else.

The same logic applies outside Google. ChatGPT, Perplexity and others build their own understanding of entities, often from the same public sources such as Wikidata and Wikipedia. Entities also matter for how AI tools match meaning rather than exact words, which I explain in embeddings and vector search. My guide on fixing AI brand misinformation covers what to do when AI tools get you wrong.

Strengthening your entity, step by step

If your entity is missing or weak, here is the order I suggest working in. Each step builds on the one before it.

1. Pick one official name and use it everywhere: site, Business Profile, social profiles, invoices, press releases.
2. Build a clear About page that states what you are, where you are, when you started and who runs the business. My About page guide walks through this.
3. Add Organization schema to your homepage or About page, with your name, logo, URL and a sameAs list of your official profiles. See my Organization schema guide.
4. Verify and complete your Google Business Profile if you serve customers locally.
5. Make your profiles match. LinkedIn, Facebook, Crunchbase, industry directories: same name, same description, same website.
6. Consider Wikidata if you meet its notability rules. Add only facts you can back with public references.
7. Earn independent mentions. Trade press, local news, podcasts and association listings confirm you are real. My guide to digital PR for SEO shows how.
8. Re-check the API every quarter and note changes to your type, description and position in results.

Example: a three-month entity plan for a local firm

Hypothetical scenario. A two-location accounting firm runs the API check and finds its brand in fourth place for its own name, typed only as “Organization,” with an old domain from before a rebrand.

• Month 1: update the About page, add Organization schema with the new domain and sameAs links, fix both Business Profiles and five main directories.
• Month 2: publish a short press release about the rebrand on a local news site and two accounting associations’ member directories.
• Month 3: re-run the API. The goal is to see the firm in first place for its name, with the new URL and a more specific type, such as AccountingService.

There is no guaranteed timeline. Google updates entities on its own schedule. But this is the work that gives it the clear signals it needs.

Fixing wrong entity data in Google

If Google has facts about you that are wrong, you have three routes:

Situation	Best route	Notes
You have a knowledge panel and can verify	Get verified, then use “Suggest edits”	Verified users get priority on panels about them
You have a panel but can’t verify yet	Use “Feedback” on the panel	Anyone can submit feedback; include public URLs that prove the correct fact
No panel, but wrong facts in AI answers or the API	Fix the sources	Update your site, schema, profiles and Wikidata; Google picks up changes over time

Google’s feedback help page explains that once you verify, Google can make some corrections directly, like links to social profiles. For bigger facts, always attach publicly accessible pages that support the change. A correction request with no proof is easy to ignore.

And remember the source table above. If Wikidata, an old directory and a press article all list your old address, fixing the panel alone won’t stick. Fix the sources too.

Questions brand owners ask about the Knowledge Graph

Can I pay Google to add my business to the Knowledge Graph?

No. There is no paid way in. Google builds entities from public sources, licensed data and verified owner feedback.

Do I need a Wikipedia page to be in the Knowledge Graph?

No. Many businesses have entities without Wikipedia articles, often supported by their Business Profile, website and other sources. Wikipedia helps, but it is not required, and you should not write your own article.

What is a kgmid and why should I save it?

It is the ID Google uses for your entity, shown in the API as @id. Saving it lets you look up the exact entity later with the ids parameter, even if a name search returns other results.

Does schema markup put me in the Knowledge Graph?

Not on its own. Schema helps Google read your official facts clearly, but it is one signal among many. Consistent profiles and independent mentions matter as much.

How is the Knowledge Graph different from how ChatGPT knows my brand?

The Knowledge Graph is Google’s structured database. ChatGPT learns from training data and web search. They often draw on the same public sources, which is why fixing your entity data helps across platforms. My post on how retrieval-augmented generation works explains the AI side.

Run your Knowledge Graph check today

This is one of the few SEO checks you can do in 15 minutes for free. Get an API key, search your brand, save your kgmid and note what Google thinks you are. Then compare it with what you actually are.

If the gap is big, start with the step-by-step list above. And if you want help building a clear, consistent entity across Google and AI search, our SEO team does this work every day. Get a free quote and we will take a look at your brand’s entity together.

For related reading in this series, see Grokipedia vs Wikipedia for brands, which covers another source AI tools now read.

Further reading on Google Knowledge Graph

• AI-Proof Keywords: Finding Queries AI Answers Can't Fully Satisfy
• What Happened to Featured Snippets in the AI Era
• Agentic AI vs Generative AI: The Difference for Marketers

Sources for this Google Knowledge Graph guide

1. Ahrefs — Google’s Knowledge Graph (May 2026)
2. Google — Introducing the Knowledge Graph: things, not strings (May 2012)
3. Google Knowledge Panel Help — How Google’s Knowledge Graph works
4. Google for Developers — Knowledge Graph Search API
5. Google for Developers — Method entities.search
6. Google for Developers — Knowledge Graph Search API usage limits
7. Google Cloud — Migrate from legacy Google Knowledge Graph Search API
8. Google Knowledge Panel Help — Submit feedback on content about you
9. Google Knowledge Panel Help — Get verified on Google
10. Google Search Central — AI features and your website
11. Google Search Central — Optimizing for generative AI features