Knowledge Graph
SEOAlso: Entity Graph · Semantic Graph
Quick definition
A knowledge graph is a structured database that maps entities (people, places, organisations, products, concepts) and the relationships between them. Search engines and AI systems use knowledge graphs to understand meaning, not just keywords, and to generate factual answers without showing a list of links.
How it varies across Australia
Australian businesses rarely audit their presence in knowledge graphs, which means entity data for local brands tends to be thinner and less accurate than for global counterparts. The gap shows up most clearly in AI-generated summaries and featured snippets where entity confidence drives inclusion.
See digital maturity and data tracking scores across Australian industries →The building blocks
Any distinct, identifiable thing: a brand, a person, a product, a location, a concept.
The typed connection between two entities. 'ACME founded by Jane Smith' is a relationship.
A property of an entity. A business has an address, a phone number, a founding year.
How certain the graph is that its data about an entity is correct, based on corroborating sources.
What it actually means
A knowledge graph is less like a spreadsheet and more like a web of facts. Google's Knowledge Graph connects billions of entities, defining what each one is and how it relates to everything else. Your business is an entity. Your industry is an entity. Your founders, your location, your products are entities. The graph knows how they connect.
When Google shows a business panel on the right side of search results, that data comes from the knowledge graph. When an AI Overview answers a factual question without citing a webpage, it's pulling from the graph. When Gemini or ChatGPT summarises your company, the accuracy of that summary depends heavily on what the underlying knowledge graph holds about you.
This is why entity SEO has become a real discipline alongside keyword SEO. Ranking for keywords is about matching text. Appearing in AI-generated answers is about being a known, trusted, well-defined entity. The two strategies overlap but they are not the same.
For most Australian businesses, the knowledge graph knows less about them than it should. Thin or inconsistent entity data across Wikipedia, Wikidata, Google Business Profile, structured data markup and authoritative directories means the graph has low confidence in your entity, which means it hedges or omits you. Fixing that is increasingly the work that determines whether AI systems mention you at all.
If the knowledge graph doesn't know who you are, the AI doesn't either.
How it shows up
The knowledge graph shows up in search and AI interfaces in several ways. The Google Business Panel (the box showing your hours, address and photos) is knowledge graph data. Featured snippets that answer 'who is X' or 'what is Y' pull from it. AI Overviews in Google Search draw on entity data when generating summaries. ChatGPT and Perplexity reference their own graph-like structures for factual grounding.
For a business, the practical signal is simple: search your brand name and see what the right-hand panel shows. If it's absent, thin or wrong, your entity confidence is low. That affects AI inclusion, featured snippets, and branded search appearance simultaneously.
The Australian context
Australia has a smaller Wikipedia and Wikidata footprint than the US or UK. Many Australian businesses and public figures that would have a Wikipedia page in larger markets do not have one here, which means the knowledge graph has fewer authoritative corroborating signals for Australian entities.
This creates a specific opportunity. Australian businesses that invest in structured data markup, consistent NAP (Name, Address, Phone number) data across directories, and even modest Wikidata entries tend to see knowledge panel improvements faster than equivalent US businesses would, simply because competition for entity confidence is lower. The ground is easier to gain.
Where people get this wrong
Related terms
Common questions
How do I get my business into Google's Knowledge Graph?
There is no direct submission. The graph builds entity confidence from corroborating signals: a complete and verified Google Business Profile, consistent NAP data across authoritative directories, structured data markup on your site, and where possible a Wikidata entry. The more independent sources agree on your entity, the faster confidence builds.
Why does my knowledge panel show wrong information?
The graph pulls from sources it trusts more than your own website. Old directory listings, outdated Wikipedia data or a conflicting Wikidata entry are common culprits. Correct the source the graph is reading from, not just your own site. The panel usually updates within a few weeks of the source being corrected.
Is the knowledge graph the same thing as a featured snippet?
No. A featured snippet is an extracted passage from a webpage. A knowledge graph answer (like the business panel or a factual one-line answer) comes from the graph's own stored entity data. Featured snippets cite a URL. Knowledge graph answers often do not.
Does the knowledge graph affect AI answers from ChatGPT or Perplexity?
Not directly. Google's Knowledge Graph is Google's own system. But ChatGPT and Perplexity have their own underlying entity data from training and retrieval. The principle holds: consistent, authoritative, corroborated entity information across the web improves how accurately any AI system describes you.
Keep exploring
About New Rebellion
New Rebellion is a marketing intelligence consultancy. We build tools, score Australian businesses on how their marketing actually performs, and publish Debrief every day. This dictionary is part of how we work in the open.
How we think →