What Is a Google Knowledge Panel, and Why Do Some People Get One?

Search your own name on Google. For some people, a box appears beside the results with their photo, a one-line role, a short bio and links to their profiles. For others there’s only a list of links. And for a few unlucky ones, there’s a box about somebody else with the same name.

That box is a knowledge panel. Google builds it automatically from what it knows about a person, place or organisation¹. Nobody writes their own.

If you’re a coach, consultant, founder, doctor, lawyer or author, your name does a lot of commercial work. Prospective clients look you up before they get in touch. A panel shows them, at a glance, that Google recognises you as a distinct, real professional. Without one, they’re left piecing you together from whatever links happen to rank.

The test Google applies has far less to do with fame than most people assume. This guide explains what a panel is, where its information comes from, how Google decides who gets one and what you can realistically do about yours.

What is a Google Knowledge Panel?

A knowledge panel is the information box Google shows when you search for something it holds in its Knowledge Graph: a person, a place, an organisation or a thing¹. On desktop it usually sits to the right of the search results. On mobile, the same information is split into one or more cards among the results².

For a person, a typical panel shows a name, a short subtitle such as “Author” or “Entrepreneur”, one or more photos, a brief description with a link to its source, a few key facts and icons linking to social profiles. We’ll take it apart piece by piece further down.

Google generates panels automatically and updates them as information on the web changes¹. It also weighs input from the person the panel is about and feedback from users. There’s no application form and no approval queue.

The Knowledge Graph: the database behind every panel

Every panel is a window onto one entry in Google’s Knowledge Graph. That’s Google’s database of things (it calls them entities) and the facts and relationships connecting them.

Google launched it on 16 May 2012 under the slogan “things, not strings”³. The idea was that Search would understand which thing you meant, rather than just matching the letters you typed. Google’s own example was “Taj Mahal”, which could mean a monument, a musician or a casino.

An entity is any single, distinct thing Google can identify and describe. You are one. So is your company, your book and your podcast. Panels belong to entities, which means that if Google doesn’t hold you as an entity, there’s nothing for a panel to describe.

At launch, the graph held more than 500 million objects and over 3.5 billion facts about them, drawn from sources including Freebase, Wikipedia and the CIA World Factbook³. Within seven months it had reportedly grown to 570 million entities and 18 billion facts, and by 2016 reports put it at 70 billion facts⁴.

The last size figure Google itself published came in May 2020: more than 500 billion facts about 5 billion entities². Anything newer comes from outside Google. Kalicube, the firm run by knowledge panel specialist Jason Barnard, estimated about 54 billion entities and 1.6 trillion facts as of around May 2024⁵. Google hasn’t confirmed that estimate, and counting methods differ, so read it as a sense of scale rather than a precise count.

One early source is worth knowing about. Google announced on 16 December 2014 that it would close Freebase, an open database it had relied on, and helped move that data to Wikidata, the free structured database run by the Wikimedia Foundation⁶. That history is one reason Wikidata still comes up whenever people talk about panels.

A short history, from 2012 to AI answers

The graph has changed a great deal since launch. In June 2018, Google opened verification to any person with a panel⁷. In July 2023, Kalicube’s tracking recorded the number of Person entities in the graph tripling in three days⁸. AI Overviews reached everyone in the US on 14 May 2024⁹, and AI Mode followed in March 2025, drawing on the Knowledge Graph¹⁰. Then, in June 2025, Kalicube recorded Google deleting more than 3 billion entities in a single week¹¹.

1. 16 May 2012

   The Knowledge Graph launches

   Google introduces “things, not strings” in the US, starting with more than 500 million objects and 3.5 billion facts.³

2. December 2012

   Rapid early growth

   The graph expands into more languages and reportedly reaches 570 million entities and 18 billion facts.⁴

3. 16 December 2014

   Freebase closes

   Google announces it will shut Freebase, one of its early sources, and helps move the data to Wikidata.⁶

4. 22 June 2018

   Verification opens up

   Google opens verification to people, organisations, sports teams, events and media properties that have panels.⁷

5. 20 May 2020

   Google’s last official size figure

   Google reports more than 500 billion facts about 5 billion entities, and says Wikipedia is one of hundreds of sources.²

6. July 2023

   A big update for people

   Kalicube’s tracking shows Person entities tripling in three days and more Writer and Author subtitles appearing.⁸

7. 14 May 2024

   AI Overviews roll out

   AI-generated summaries reach everyone in the US, backed by what Google calls a knowledge base of billions of facts.⁹

8. 5 March 2025

   AI Mode arrives

   Google announces AI Mode and names the Knowledge Graph among the real-time sources it draws on.¹⁰

9. June 2025

   The clarity cleanup

   Kalicube records the graph shrinking by 6.26%, with more than 3 billion entities removed in one week.¹¹

How Google decides, and why you might not have one yet

Google doesn’t publish an eligibility checklist. Between what Google has said and what practitioners have observed over many years, though, the process is reasonably well understood. It runs roughly like this.

1. 1

   Google holds you as an entity

   Panels are tied to entities in the Knowledge Graph. If Google hasn’t recognised you as a distinct entity, there is nothing for a panel to describe.

2. 2

   It gathers facts from many sources

   Your website, your profiles, press coverage, open databases such as Wikidata, licensed data and structured markup all feed in. Wikipedia is only one source among many.

3. 3

   It works out that they all describe one person

   Google has to match your site, your LinkedIn, a podcast page and a news article to the same individual. Consistent names, roles and descriptions make that easy. Contradictions make it hard.

4. 4

   It has to be confident

   A panel appears when Google is sure of the facts and judges that a short factual summary will help the person searching.

5. 5

   The search has to be about you

   The query needs to point clearly at one entity. A distinctive name, or a name plus a profession or a city, makes that easier.

6. 6

   It keeps checking

   Panels update automatically as the web changes, and large Knowledge Graph updates can add, reshape or remove entities in bulk.

The first, second and last of those steps come straight from Google. Panels are tied to Knowledge Graph entities and change automatically as information on the web changes¹. Wikipedia is one of hundreds of sources, alongside material shared across the web, open databases, licensed data for areas such as music and sport, and structured markup from site owners².

The middle steps are practitioner explanation. Barnard frames the decision as confidence plus usefulness: Google shows a panel when it trusts the facts and believes a factual summary helps the searcher¹². That fits Google’s own stated aim of showing the most relevant and popular information about a topic².

Step five is simply how search works. “John Smith” could be thousands of people. “Dr Priya Raman, cardiologist, Leeds” points to one. If you share a common name, the context around it does a lot of the work.

Put all of that together and the picture is fairly consistent. Google isn’t asking whether you’re famous. It’s asking whether it’s sure who you are. Consistent with that, Google’s 2018 verification update set no celebrity threshold⁷.

Why you might not have one (or have the wrong one)

If you’ve searched your name and found nothing, or found a panel that isn’t quite you, one of these is usually the reason.

• Google doesn’t see you as a distinct entity yet. Your information exists, but it’s scattered and nothing ties it together.
• Your details don’t agree. One profile says “consultant”, another says “managing director”, your old speaker bio lists a job you left years ago. Practitioners most often point to inconsistency like this as the reason panels fail to appear¹².
• Nobody else is vouching for you. If the only pages describing you are pages you wrote, Google has little to corroborate them against.
• You share a name. Google may blend your details with a namesake’s, or show the more prominent person’s panel when someone searches your name.
• Google tidied you away. Large Knowledge Graph updates add, retype and delete entities in bulk. Kalicube’s tracking suggests major updates can affect 60 to 80% of entities, with smaller shifts every two to three weeks⁸. Google hasn’t confirmed those figures, but they explain why panels sometimes vanish while nothing changed on your side.

The June 2025 “clarity cleanup” shows how sharp those swings can be. After a year in which the graph grew by 2.79%, Kalicube recorded it shrinking by 6.26% across updates on 13 and 20 June, with some entity types cut far harder than others¹¹.

Five myths that stop experts from trying

Most of the professionals who assume a panel isn’t for them are working from one of these beliefs.

Google’s 2018 update covered organisations, sports teams, events and media properties as well as people⁷. Panels follow Knowledge Graph entities, not celebrity¹.

Google said as much in 2020². Kalicube’s tracking found that Wikipedia’s share as the triggering source for Person entities fell from 40.9% in June 2023 to 12.44% in September 2023⁸. And please don’t write a Wikipedia article about yourself. Barnard calls creating a page that doesn’t already exist generally a bad idea, because Wikipedia’s notability and conflict-of-interest rules mean self-created pages are often deleted¹².

Google says panels are generated automatically and that, under its current policy, it doesn’t manually create or delete them¹³. Its Search documentation adds that Google doesn’t accept payment to crawl a site more often or rank it higher¹⁴.

Even verified people can’t write a custom subtitle or description, reorder their social profiles or add a photo where none exists¹³.

Google’s structured data documentation is explicit that markup doesn’t guarantee that the features using it will show up¹⁵. Treat markup as one clear signal among many, not a switch.

The anatomy of a person’s panel

Knowing where each part of a panel comes from tells you which parts you can influence and which you can’t. Here’s a typical panel for a person, labelled.

The table below sets out where each part usually comes from and how much say a verified person has over it.

Two parts need more explanation. The description is the one people most want to change, and it’s the one you can’t edit directly. When it isn’t drawn from Wikipedia, practitioners commonly see the person’s own About page quoted¹², which is a strong argument for making that page clear and factual.

The subtitle is the other. It’s generated from how Google understands your main role, and it can shift during updates. In Kalicube’s tracking of the July 2023 update, the share of Person entities with a subtitle fell from 57.1% to 16.58%, while the number carrying “Writer” or “Author” subtitles rose by 21%⁸. The practical lever is describing your role in the same words everywhere.

Claiming your panel: what verification gives you

Verification is for people who already have a panel. You can’t claim a panel that doesn’t exist, and not every panel can be claimed straight away. Google suggests checking back periodically for the option¹⁶.

When the option is there, the process is short:

1. Sign in to a Google Account.
2. Search your name and find your panel.
3. Click “Claim this knowledge panel” at the bottom of the panel (it sometimes sits under the “About” section).
4. Review the information Google shows you.
5. Sign in to an official profile Google already associates with you. The options Google lists are YouTube, Search Console, X (formerly Twitter) and Facebook. If Google can’t find a linked profile, it asks for more information to confirm who you are¹⁶.
6. Once you’re verified, you can add other people as owners, managers or contributors.

After that, a “Suggest edits” option appears on your panel when you’re signed in with the verified account and Web & App Activity is turned on. Flag each fact separately, describe the change and link to publicly accessible pages that confirm it. Google says it reviews verified feedback “within a few days”, though it can take longer¹³.

Here’s what verification does and doesn’t let you do.

One caveat from practice: an edit Google accepts today can be undone later if the rest of the web keeps saying something different¹². Correct the panel, then correct the sources too. The full walkthrough, including how to word an edit request, is in How to Claim and Edit Your Knowledge Panel.

Knowledge panels and AI answers

The Knowledge Graph that powers your panel is also one of the fact sources behind Google’s AI features.

What Google has confirmed

AI Overviews rolled out to everyone in the US on 14 May 2024, with Google aiming to reach more than a billion people by the end of that year. Google described them as backed by a knowledge base of billions of facts about people, places and things⁹.

Google’s own explainer says AI Overviews use a customised Gemini model that works alongside its Search systems, including the Knowledge Graph¹⁷. When Google announced AI Mode, it said the feature taps “fresh, real-time sources like the Knowledge Graph”¹⁰.

Google also says there’s no special markup or extra requirement for appearing in these AI features. A page needs to be indexed and eligible to show a snippet in normal Search¹⁸.

So being a clear, well-understood entity in Google’s graph plausibly helps Google’s AI describe you accurately. It doesn’t guarantee that an AI answer will mention you.

What’s interpretation, not fact

Barnard argues that the June 2025 cleanup was intended to give AI Overviews and AI Mode a leaner, higher-confidence dataset¹¹. It’s a plausible reading. Google hasn’t said so.

What’s overstated

Some guides claim your Google panel feeds ChatGPT or Microsoft Copilot. There’s no evidence that OpenAI or Microsoft have access to Google’s Knowledge Graph. Those systems build their own understanding of people from web data, their own search indexes and partners.

What is true is that the habits that earn a panel (a consistent bio, independent sources that agree, accurate structured data) are things every system reading the open web can pick up. That’s a reasonable expectation, not a measured result. No published study has shown that having a panel causes more AI mentions. For what OpenAI does say about how ChatGPT picks names, see How ChatGPT Decides Which Coaches to Recommend.

The practical path to a panel

You can’t apply for a panel, but you can make it easy for Google to understand you. It starts with knowing where you stand.

Start by checking where you stand

Search your name in a private browser window, so your own history doesn’t colour the results, and check on both desktop and mobile. Then try your name with your profession, and your name with your city.

As you look, ask four questions. Is there a panel? Is it about you? Is the information right? Is someone else’s panel showing up for your name?

Separately, Google’s public Knowledge Graph Search API returns the entities Google holds for a name, along with their IDs and types such as Person. Google notes that the API isn’t suitable as a production-critical service¹⁹, so treat it as a useful check, not the final word on what Google knows.

Build the signals, step by step

1. 1

   Choose one home page for your identity

   Make your About page, or your homepage, the definitive statement of who you are. Avoid relying on a social profile or an employer’s page as the main source.

2. 2

   Write one clear, factual bio

   Open with “[Name] is a [role] who [does what] for [whom].” Put the present first and work back. List credentials, books, awards and affiliations. Keep a master version and reuse it.

3. 3

   Use the same details everywhere

   Same name, title, photo and short bio on LinkedIn, X, YouTube, speaker pages, podcast guest pages, association directories and your firm’s team page.

4. 4

   Link both ways

   Every profile should link to your home page, and your home page should link out to every genuine profile.

5. 5

   Add Person structured data

   Mark up your home page with JSON-LD so machines can read who you are. There is an example below.

6. 6

   Get your home page indexed

   Submit it in Google Search Console so Google finds it sooner.

7. 7

   Earn independent mentions

   Interviews, podcast appearances, conference speaker pages, bylined articles, awards and memberships. Pages mostly about you, written by someone else, carry the most weight. Quality beats volume.

8. 8

   Use the directories that matter in your field

   The medical register for doctors, bar listings for lawyers, Crunchbase for founders, Goodreads or Amazon Author Central for authors. Skip mass directory submissions.

9. 9

   Claim it when it appears, then keep it current

   Verify your panel, keep the facts up to date and check it regularly.

These steps follow the method practitioners have refined over years, most fully set out in Barnard’s knowledge panel checklist¹². Two points from that work are worth underlining. Inconsistency is the blocker practitioners cite most, so step three matters more than it looks. And piling up low-quality directory profiles tends to work against you, so a few credible sources in your field beat dozens of spammy ones¹².

Add Person markup to your home page

Structured data is a small block of code that describes you in a format machines read reliably. Google recommends the JSON-LD format for structured data in general²¹. Here’s a complete example for a fictional coach, Jordan Ellis. Every name and URL in it is a placeholder. If the idea is new to you, Structured Data in Plain English explains it from the start.

{
  "@context": "https://schema.org",
  "@type": "Person",
  "@id": "https://www.example.com/about#person",
  "name": "Jordan Ellis",
  "alternateName": "Jordan A. Ellis",
  "jobTitle": "Executive Leadership Coach",
  "description": "Jordan Ellis is an executive leadership coach who helps first-time founders build and lead their teams, and the author of The First Hundred Hires (2025).",
  "url": "https://www.example.com/about",
  "image": "https://www.example.com/images/jordan-ellis.jpg",
  "worksFor": {
    "@type": "Organization",
    "name": "Ellis Leadership Partners",
    "url": "https://www.example.com"
  },
  "alumniOf": {
    "@type": "CollegeOrUniversity",
    "name": "Example State University"
  },
  "knowsAbout": [
    "Executive coaching",
    "Leadership development",
    "Startup hiring"
  ],
  "sameAs": [
    "https://www.linkedin.com/in/example-jordan-ellis",
    "https://www.youtube.com/@examplejordanellis",
    "https://x.com/example_jellis"
  ],
  "subjectOf": {
    "@type": "Article",
    "headline": "Interview: Jordan Ellis on hiring your first team",
    "url": "https://news.example.org/interview-jordan-ellis"
  }
}

On your site, this goes inside a <script type="application/ld+json"> tag on your About page. Here’s what each property does, using the definitions on schema.org²²:

• @context and @type say this uses the schema.org vocabulary and describes a Person. @id gives that person a stable identifier other markup on your site can refer to.
• name and alternateName state your name and any variant people use, so Google can connect mentions of either.
• jobTitle states your role in the same words you use everywhere else.
• description is your one-sentence bio, matching the master version on the page.
• url and image point to your home page and a clear, current headshot.
• worksFor names your firm or employer as an Organization with its own URL.
• alumniOf and knowsAbout add your education and areas of expertise.
• sameAs lists your genuine profiles on other sites. schema.org defines it as a reference page that unambiguously indicates the item’s identity²³. List only profiles that really are yours¹². If you have a legitimate Wikidata item, its URL belongs here too.
• subjectOf points to an article about you, such as an interview, which ties independent coverage back to you.

Check the code at validator.schema.org. Barnard recommends it over Google’s Rich Results Test, which only checks types that produce rich results¹². And keep expectations straight: markup helps Google understand you, but Google doesn’t guarantee it will change how you appear¹⁵.

How long it takes

There’s no official timeline, and anyone who gives you a firm date is guessing. Practitioner estimates vary widely. Barnard’s manual puts it anywhere from a few days to a year¹², and Kalicube’s site suggests three weeks to six months for verification²⁴. The clearer and better corroborated your presence already is, the shorter the wait tends to be.

Where to start

A knowledge panel is Google’s summary of what it’s confident it knows about you. Fame has very little to do with it. Clarity, consistency and independent corroboration have a great deal to do with it, and those are things an established expert can build.

Start by finding out what Google sees today. Run your name through the free Knowledge Panel Checker. Then compare your bio, job title and photo across every profile you own, and fix anything that disagrees. If you’d like a second pair of eyes on what’s helping and what’s holding you back, book a free audit.

Keep reading

Knowledge PanelsHow to Claim and Edit Your Knowledge PanelShobin K Jose · 22 Sep 2026 · 17 min readGuidesThe Name Search Checklist: 12 Things Prospects See Before They CallShobin K Jose · 6 Oct 2026 · 18 min readAI searchHow ChatGPT Decides Which Coaches to RecommendShobin K Jose · 9 Oct 2026 · 14 min read