The Tin Man Problem
The Tin Man had a heart all along. He just needed someone to recognise it.
Your brand has the same problem, only it is not a wizard you need. It is a knowledge graph.
Every brand has soul. A point of view, a way of doing things, a reason people choose you over the next name on the list. It is there in the way you talk to clients, in the methods you have developed, in the language you use, in the philosophy you would describe over a coffee but have never thought to write down in machine-readable form.
And that is the problem. AI cannot see it.
Not because AI is broken. Not because your content is poor. Because AI does not read your website the way a human does. It does not absorb your tone, appreciate your story, or intuit the things that make you distinctive. It reads structured data. It resolves entities. It reasons from what has been formally declared. And if all you have declared is what type of thing you are such as: a consultant, an educator, a software company then that is all any AI will ever say about you.
Factually correct. Personality-free. One of the herd.
Go On. Ask.
Open ChatGPT, Gemini, Perplexity, or whichever one your customers are already using instead of Google. Type your name. Type your company name. Read what comes back.
If it sounds like it could describe any of your competitors, you are looking at the Tin Man problem in action. Your brand has a heart. AI just returned a description of the tin.
This is not hallucination. The AI is not making things up. It is not getting you wrong in some dramatic, catchable way. It is doing something quieter and more damaging: it is getting you approximately right.
Approximately right means generic. It means the description of your brand could be swapped with a dozen others and nobody would notice. The facts are there: name, role, industry, maybe a product or two, but the character is gone. The philosophy is gone. The thing your clients actually buy, which is almost never a commodity, has been reduced to one.
Why the Heart Is Invisible
Schema.org is the structured data standard that search engines and AI agents rely on. It was built for classification. It is exceptionally good at telling machines what you are. A Person. An Organization. A Product. It records facts: name, role, location, affiliation.
What it does not do, as almost everyone implements it, is encode who you are. Your teaching philosophy. Your approach to client relationships. The language you have coined. The methods you have developed. The reason someone hires you instead of the other seven people who do the same thing on paper.
That layer, let’s call it the soul of the brand, exists everywhere in the human experience of your business. It is in your conversations, your content, your reputation. It is real and it is valuable. But it has never been translated into structured data, because nobody thought to ask.
The result is a kind of digital homogenisation. Thousands of brands, all correctly classified, all properly validated, all described by AI in language so generic it could have been written by committee. The schema was built for classification. Nobody built it for character.
Until now.
What Soul Sounds Like
There is a mathematician called James Tanton. If you are not familiar with his work, the short version is this: he teaches mathematics as a joyful, story-driven, exploratory experience. His signature creation, Exploding Dots, has reached millions of learners worldwide. He coins language, terms like: Twelvety, The Flail, because he believes playful, accessible vocabulary is how you invite people into mathematical thinking rather than gate-keep them out.
His teaching philosophy is his brand. Inseparable. Distinctive. Immediately recognisable to anyone who has spent five minutes with his work.
Ask an AI about James Tanton without a knowledge graph and you get something like: “James Tanton is a mathematician and educator known for making mathematics accessible.”
Factually correct. Could describe a hundred people. The soul is missing. The Tin Man has been described without his heart.
Now consider what happens when the structured data actually encodes who he is, not just the biographical facts, but the teaching philosophy, the pedagogical approach, the audience relationship, the characteristic language, the intellectual methods, the reasons behind the methods. When an AI agent traverses that knowledge graph, it does not return a generic summary. It returns something that feels authentically like James Tanton, because the meaning it drew from was authentically his.
The heart was always there. The knowledge graph simply made it visible to machines.
Why More Content Will Not Fix This
The instinct, when brands discover they sound generic in AI, is to write more. Better blog posts. Richer bios. More detailed about pages. Surely if AI had more to read, it would capture the nuance. It will not.
More content gives AI more raw material, but raw material is not structure. An AI agent does not read your blog archive and synthesise a personality. It looks for declared, structured, machine-readable signals — and if those signals only declare facts, it will only return facts. More facts, perhaps. Better-sourced facts. But still just facts.
This is also not a prompt engineering problem. You cannot fix brand representation by writing a clever system prompt for someone else’s AI. You do not control ChatGPT. You do not control Perplexity. You do not control whatever agentic workflow your potential client is running next year. The only thing you control is the structured data on your own domain — and that is exactly where the solution lives.
Giving AI Your Heart
What VISEON builds is an intent layer within a Schema.org knowledge graph. It sits above the factual graph, the names, roles, awards, publications that standard schema already handles, and encodes the things that make a brand distinctive:
Voice and tone. Not a style guide buried in a PDF, but structured declarations that any AI agent can read. How the brand communicates. What register it uses. What it sounds like when it is being itself.
Philosophy and method. The principles behind the work. The approach that shapes every engagement. The reason the brand exists beyond commercial exchange.
Coined language. The terminology the brand has created or adopted as its own. The vocabulary that signals expertise and identity simultaneously.
Audience relationship. How the brand relates to the people it serves. Teacher and student. Advisor and client. Collaborator and peer. Each relationship implies a different mode of communication, and AI needs to know which one applies.
Characteristic actions. Not just what the brand offers, but how it delivers. The patterns, methods, and behaviours that clients experience and competitors do not replicate.
Each of these is encoded as structured data: machine-readable, source-attributed, and traversable by any AI agent that encounters it. Not locked inside a single platform. Not dependent on a prompt. Published once on your domain and available everywhere.
The Herd, or the Heart
AI is becoming the front door to every brand. The question is no longer whether your customers will ask AI about you, because they already are. The question is what AI says when they do.
If your structured data encodes only facts, AI will return only facts. You will sound like everyone else who shares your job title, your industry, your category. You will be findable, perhaps, but indistinguishable. Part of the herd.
If your structured data encodes your soul: your philosophy, your voice, your methods, your language, then AI has something to work with. Something distinctive. Something that represents the brand as it actually is, not as a generic approximation of what it might be.
The Tin Man had a heart all along. He just needed someone to make it visible.
Your brand is no different.
What does AI say about you? Start with a Knowledge Graph Assessment →
Summary
This is a sophisticated argument for moving from Technical SEO — making sure you can be crawled — to Semantic Branding: making sure you can be understood. If you want to be more than a commodity in a generated summary, you have to speak the language AI uses to categorise the world.
— Gemini, Google DeepMind, on reading this article
Frequently Asked Questions
What does AI say about my brand?
If you ask ChatGPT, Gemini, Perplexity, or any AI agent about your brand and the response sounds like it could describe any of your competitors, you have a structural problem. AI agents do not read websites the way humans do. They read structured data, resolve entities, and reason from what has been formally declared. If your Schema.org markup only encodes facts (name, role, industry, location) then AI will return only facts. The result is a description that is factually correct but personality-free. VISEON calls this the Tin Man problem: your brand has a heart, but AI cannot see it because it has never been encoded in machine-readable form. A Knowledge Graph Assessment reveals the gap between what AI currently says about you and what it should say.
Why does my brand sound generic in AI?
Schema.org was built for classification, not character. It tells machines what you are (a Person, an Organization, a Product) but not who you are: your philosophy, your voice, your methods, or the reason someone chooses you over seven competitors who look the same on paper. Almost every Schema.org implementation stops at this factual layer. The result is digital homogenisation: thousands of brands, all correctly classified, all properly validated, all described by AI in language so generic it could have been written by committee. Your brand has soul. It is in your conversations, your content, your reputation. But it has never been translated into structured data, because nobody thought to ask. That is what VISEON changes.
What is an intent layer in a knowledge graph?
An intent layer is a set of structured declarations that sit above the factual graph (names, roles, awards, publications) and encode the things that make a brand distinctive. It includes voice and tone (how the brand communicates), philosophy and method (the principles behind the work), coined language (terminology the brand has created), audience relationship (teacher-student, advisor-client, collaborator-peer), and characteristic actions (the patterns and behaviours clients experience). Each declaration is machine-readable, source-attributed, and traversable by any AI agent. VISEON builds intent layers using Schema.org PropertyValue nodes, making brand personality available to every AI system that reads the knowledge graph, not just systems the brand controls.
Can I control how AI describes my brand?
Not through prompt engineering, and not by writing system prompts for someone else’s AI. You do not control ChatGPT, Gemini, Perplexity, or whatever agentic workflow your potential client uses next year. The only thing you control is the structured data on your own domain. If that structured data encodes only facts, AI will return only facts. If it encodes your philosophy, voice, methods, and distinctive character through an intent layer in a Schema.org knowledge graph, then AI has something meaningful to work with. The solution is not to try to influence other people’s AI systems. It is to publish rich, intent-bearing structured data once on your domain and let every AI agent that encounters it represent you accurately.
What is Semantic Brand Personalisation?
Semantic Brand Personalisation is the practice of encoding brand identity (voice, philosophy, methods, coined language, audience relationships) into a Schema.org knowledge graph so that AI agents can represent the brand accurately and distinctively. It moves beyond Technical SEO (making sure you can be crawled) to Semantic Branding (making sure you can be understood). Standard schema tells AI what type of thing you are. Semantic Brand Personalisation tells AI who you are. VISEON implements this through an intent layer of PropertyValue nodes, DefinedTermSets for coined terminology, and typed Action nodes for characteristic behaviours, creating a machine-readable digital twin that any AI agent can traverse.
What is digital homogenisation?
Digital homogenisation is the effect of AI describing every brand in generic, interchangeable language because the structured data available encodes only classification (Person, Organization, Product) and not character. When thousands of brands declare the same entity types with the same factual properties and no distinguishing semantic signals, AI has no basis for differentiation. Every consultant sounds the same. Every educator sounds the same. Every software company sounds the same. The descriptions are factually correct and completely forgettable. Digital homogenisation is distinct from digital obscurity (where a brand is invisible to AI entirely). A homogenised brand is findable but indistinguishable. The remedy is encoding distinctive brand signals (philosophy, voice, methods, coined language) directly into the knowledge graph through an intent layer.