Google's Trusted Sources for Knowledge Panel Attributes
Google shows attributes for an entity in its Knowledge Panel. But one thing we don’t often think about is “where is Google sourcing this information?”. At Kalicube we have figured that out !
Historical (and detailed) data is available as part of the Kalicube Pro SaaS platform
Knowledge Panel: Google's Public Declaration of Confidence
The information box Google shows when it is confident it knows who or what you are.
A Knowledge Panel is the information box that appears on the right side of Google's search results (desktop) or at the top (mobile) when someone searches for a known entity — a person, company, product, place, or concept. It draws its data from Google's Knowledge Graph and represents Google's confident, structured understanding of that entity.
Getting a Knowledge Panel is not a design decision — it is an algorithmic one. Google shows a panel when it has enough verified, corroborated information to be confident about an entity's identity. The presence of a Knowledge Panel signals that you have passed the critical Understandability threshold in the AI Engine Pipeline.
A Knowledge Panel also feeds directly into how AI assistants talk about you. ChatGPT, Perplexity, and Google's own AI Mode all reference Knowledge Graph data when forming answers. A brand with a Knowledge Panel is one Google is willing to vouch for — and AI platforms treat that vouching as a trust signal.
The tools on this platform — Knowledge Graph Explorer, Schema generators, Trusted Sources — all contribute directly to triggering and maintaining a Knowledge Panel.
The Algorithmic Trinity
The three sources every AI assistant draws on when talking about your brand.
When an AI assistant — ChatGPT, Google AI, Perplexity, Copilot, Claude — forms an opinion about your brand, it draws on three independent sources simultaneously.
- What it was trained on (LLM Memory). Facts baked into the model during training. Reliable for well-known brands, unreliable for everyone else — and always at risk of being outdated.
- The trusted encyclopaedia (Knowledge Graph). Structured, verified fact databases that AI platforms cross-reference when they want to state something as confirmed fact rather than a guess. Getting into the Knowledge Graph is the difference between AI saying you "are" something versus "claims to be."
- What it finds right now (Search Index). Live web content queried at the moment of response. Clear, crawlable, entity-accurate pages feed this source directly.
Optimising for one source while ignoring the other two means two out of three minds are working against you. The Kalicube Process aligns all three.
Entity Home: Your Canonical Reference Page
The single page that is the machine's canonical reference point for facts about you.
An Entity Home is the one page that a machine treats as the definitive, authoritative source of facts about an entity — a person, company, product, podcast, or any other named thing. It is not just a homepage. It is the specific URL that Google (and other AI platforms) reconcile all other mentions back to.
When Google sees your name mentioned on 500 different websites, it needs a single reference point to confirm which entity all those mentions refer to. The Entity Home is that reference point. It is where facts about the entity live, where structured data (Schema markup) is most impactful, and where the entity's relationship to the Knowledge Graph is anchored.
Getting your Entity Home right is the single highest-leverage action in the Kalicube Process. It is the input you control most completely. The Schema tools on this page generate the structured data that makes your Entity Home machine-readable.
isReference: The Domains Google Uses to Validate Entity Existence
The small fraction of domains Google trusts enough to use as evidence when validating an entity.
Not all websites are equal in Google's eyes. When Google decides whether an entity deserves a place in its Knowledge Graph, it looks for corroboration from sources it has designated as trustworthy references. These are called isReference domains — sites Google uses as evidence that an entity is real, notable, and accurately described.
The Trusted Sources tool on this platform shows which domains carry isReference status. Only about 3.65% of all domains monitored qualify. A mention on one of these domains is not just a backlink — it is a signal to Google that your entity has been independently verified by a source it trusts.
Getting your entity mentioned accurately on approximately 30 isReference URLs is the practical path to a stable Knowledge Panel for most brands. Wikipedia is one such source, but it covers only 0.012% of the Knowledge Graph. For most brands, the real opportunities lie in niche-specific directories, professional associations, certification bodies, and industry publications that are specific to your field.
This concept was confirmed by the Google API leak of 2024, which showed that Google explicitly tracks which pages reference which entities using an isReference property.
Why Knowledge Panel Attribute Sources Matter for Google, ChatGPT, and Every AI Platform
Google's Knowledge Panel is built from verified attribute data pulled from specific trusted sources. The same sources feed directly into how ChatGPT, Google AI Mode, Perplexity, and Microsoft Copilot describe your entity when users ask about you. Understanding where each attribute comes from — and which source Google trusts most — lets you target your effort precisely rather than guessing.
1. LLM Memory
LLMs like ChatGPT learned from the same third-party sources and Google products shown in this chart. Attributes that appear consistently across multiple source types are the ones AI states with confidence. Attributes from a single source get hedged.
2. Knowledge Graph ← THIS TOOL
This chart shows the source breakdown for your Knowledge Panel attributes — the exact data Google has stored about your entity. Incorrect attributes here propagate directly into every AI platform's understanding of you.
3. Search Index
WebFacts and WebFacts 2nd Gen attributes are deduced from web content. What your pages say, how structured your Schema markup is, and which trusted sources reference you all feed this category directly.
An Incorrect Attribute Here Means an Incorrect Answer in ChatGPT
If Google's Knowledge Graph has the wrong job title, the wrong founding date, or the wrong location for your entity, that error appears in your Knowledge Panel — and is cited as verified fact by Google AI, ChatGPT, Perplexity, and Copilot. You cannot correct an AI hallucination by arguing with the AI. You correct it by fixing the source data.
Kalicube Pro identifies exactly which source is feeding each incorrect attribute and runs the targeted process to correct it — at the Knowledge Graph level, so the fix propagates to all AI platforms automatically.
We track where Google is getting the Knowledge Panel attributes from. Although Google corroborates the information it shows on multiple authoritative, trustworthy third party sources, we can identify what the initial source is - incredibly valuable information for anyone wanting to ensure the accuracy of the information Google shows about their Entity (brand, person, film, product…)
The actual root sources for these attributes is no longer a mystery: the distinction between the sources Google is using is clear to us at Kalicube.
What Sources Does Google Use for Knowledge Panel Attributes?
The titles of the sources are ours, not Google's. But the distinction between them is 100% clear.
- 3rd Party sources – curated by humans such as Wikipedia and Wikidata.
- Google sources – curated by humans through Google products such as Google Business and Google Podcasts.
- WebFacts – deduced by Google’s Knowledge Graph algorithms from the information taken from a seed set of trusted sources.
- WebFacts 2nd Gen. – deduced by Google’s Knowledge Graph algorithms from the information in Google’s web index. This is the key to the future of Knowledge Panel management that we are aiming to master at Kalicube SAS.
Download your free copy of Kalicube's definitive guides
Guide to SEO for the AI Era
Knowledge Panel Instruction Manual
The AI salesforce is already working.
Find out if it’s working for you.
25 billion data points across 15 sources. 11 years of longitudinal tracking. The Kalicube AI Authority Index uses all of it to audit your brand across 9 AI platforms in 60 seconds — free, no login required.
Run your free AI audit → Works for people, agencies, and companies