CORE Knowledge Framework
The 3 Truth Layers AI Uses to Describe Your Business
Why your website alone is not enough and what to build instead.
A few months ago I started running a weird little test.
I would type a company name into LLMtel.com or directly into an AI tool like ChatGPT and ask a simple question: “What do they do?”
The results were all over the place. Some companies were described perfectly. Others were described like a stranger who only half-listened at a cocktail party.
The pattern was obvious: the AI was not “reading their website.” It was reading their identity across the internet.
That is when I stopped thinking about SEO as “ranking pages” and started thinking about it as something more fundamental: truth.
Truth comes in layers
When an AI system answers a question about your business, it is usually pulling from three layers of information:
· Your website (what you say about yourself).
· Your structured data and on-site relationships (how you define yourself to machines).
· The public knowledge graph (what the world says about you, and how systems model you).
The three layers that AI uses to model your business.
To make this easy to teach (and easy to implement), I named it the CORE Knowledge Framework.
CORE = Content, Ontology, Representation, Entity
Here is the simplest way to remember it:
C = Content (Layer 1): your human source of truth. Your website as the canonical, authoritative source.
O = Ontology (Layer 2): structured data + relationships. Machine-readable truth.
R = Representation (Layer 3): public knowledge graphs. How Google and AI systems represent you.
E = Entity: the through-line. Your brand, people, products, and services must remain consistent everywhere.
If you are thinking, “Okay… but what does that actually mean for me?”, keep reading. This gets practical fast.
C = Content: Your source of truth for humans
Layer 1 is the part everyone understands. It is the marketing site. The “pretty” website. The one designed to persuade and convert.
And yes, it still matters. People still need to trust you. People still want proof.
· Clear positioning (what you do, who it is for, why you).
· Proof (case studies, testimonials, outcomes).
· Clarity (UX, navigation, obvious next steps).
But here is the twist: machines do not experience your website the way humans do.
O = Ontology: Your source of truth for machines
Machines extract facts. They do not “feel” your brand. They categorize you.
This is Layer 2 and Ontology is how you make that categorization accurate. It is the structured representation of:
· Entities (your company, your people, your locations, your services).
· Attributes (names, titles, addresses, categories).
· Relationships (who leads what, what service belongs to what brand, what location belongs to what company).
Get Dave Burnett’s stories in your inbox
Join Medium for free to get updates from this writer.
This is why schema markup matters. Not because it is trendy, but because it is one of the clearest ways to say: “Here are the facts. Here is how they connect.”
A quick example (why structure beats fluff)
If I write “We are a world-class agency delivering innovative solutions,” a human might roll their eyes and keep reading.
A machine gets almost nothing from that.
But if I structure facts like this:
· Organization: AOK Marketing (Digital Agency)
· Locations: Toronto, New York
· Leadership: Dave Burnett (CEO)
· Services: AI search optimization, SEO, paid media
Now the machine can place you in the right category and relate you to the right concepts.
R = Representation: The public truth AI builds about you
Layer 3 is where most people get blindsided.
You can build a great website and even add structured data, but if the public web contradicts you, the AI has a problem. It has to choose which sources to trust.
Representation comes from signals like:
· Your LinkedIn company page and leadership profiles
· Google Business Profile
· Directories and association listings
· Reviews and third-party mentions
· Who you are commonly compared to (“people also search for”)
You are not building this layer. You are shaping how you are represented.
E = Entity: The through-line (the part that makes it all work)
Entity consistency is boring. It is also the reason some companies get described correctly and others do not.
Here is what breaks entity consistency:
· Three different versions of your company name across the web.
· Old addresses on directories you forgot existed.
· Leadership bios that do not match reality.
· Services listed on LinkedIn that are not on your website (or vice versa).
When that happens, the model tries to reconcile the mess. Sometimes it can. Sometimes it cannot. Either way, you are not in control of the output.
A simple way to implement CORE
If you want a practical starting point, here is a simple sequence:
1. Write your Brand Factbook (one page of stable facts).
2. Make your website the best human explanation of those facts (Content).
3. Publish the structured version of those facts (Ontology).
4. Clean up the public confirmations of those facts (Representation).
That is it. That is the framework.
The real goal
The goal is not “trick the AI.” The goal is to make the AI confident in you and your services.
When your content, structure, and public signals align, you stop being a guessing game. You become an entity the system can trust and recommend.
In other words: build your CORE.
