Pasquale Caiazzo
AI systems already tell people who you are before they ever visit your site. The only question is whether that version was authored by you, or invented by the model.
Readable, verifiable, citable.
I'm Pasquale Caiazzo, Semantic Entity Architect. I design the informational layer that represents a site: structured data, verifiable entities, knowledge graphs that tell an AI system who you are and what you do — the infrastructure that determines whether a model reads you with precision or fills the gaps with guesswork.
The starting point is always the same: what do ChatGPT, Claude, and Gemini know about you right now? What's accurate, what's missing, what do they invent? From there, I build a presence that machines can cite with confidence, rather than reconstruct from noise.
For years I optimised messages for people: advertising, CRM, digital marketing. Then I changed my audience. Today I don't work to make an AI sell better, but to make it understand precisely. In a web increasingly mediated by agents, that's where it's decided who gets found and who disappears.
The markup is a technicality. What to say about you, and the sources that back it up, is not. The craft isn't in the code: it's in the deliberate mix of information you choose to share.
What it means, in practice, to build entities instead of content
Four ways semantic architecture becomes concrete work: from a line of JSON-LD to an entity on Wikidata, to the knowledge bases that power an AI assistant without hallucinations.
01 / Structured Data (JSON-LD)
Custom Schema.org markup, not templates. JSON-LD data that tells a model who you are, what you offer, and which verifiable entities you're connected to. So your identity is read, not guessed.
02 / Wikidata Entities
Connection to a verified, disambiguated entity in the most authoritative open registry — and, where there is sufficient notability, a dedicated entry. When an LLM looks you up, it finds an authoritative anchor instead of an approximation assembled from web noise.
03 / RAG Knowledge Bases
Manuals, catalogues, procedures, archives: internal knowledge reorganised into formats a RAG system can query. An answer engine drawing from verified sources, not generic probabilities.
04 / Semantic Automation
Pipelines that keep the infrastructure alive: structured data updates when content changes, entities stay consistent, and continuous monitoring shows how major AI systems describe you over time.
Interactive Demo: The Semantic Graph
As an experimental entity-linking case study, this graph maps the historical evolution of linked data and semantic standards. Drag the nodes and explore the connections to see how information is structured for machines.
Drag the graph to pan, scroll to zoom. Click a node to view its historical details and semantic relations on Wikidata.
The problem changes. The method doesn't.
I work with professionals who want a precise semantic identity, with businesses that want to be present in answer engines, and with impact organisations that want their mission told for what it actually is. The method doesn't change. What changes is the starting point, and what's at stake.
For Professionals
See the serviceYour reputation, today, is summarised in two lines by a language model for anyone who looks you up. I build your machine-readable entity — JSON-LD on your site and the channels you control, linked to verifiable Wikidata references — so ChatGPT, Claude, and Gemini describe your expertise accurately, with minimal hallucinations.
For SMEs
See the serviceAnswer engines are becoming the first point of contact between a business and its customers. I map structure, products, and services into structured data and knowledge graphs, so the company is found and cited where people ask purchasing questions.
For Nonprofits & B Corps
See the serviceA nonprofit or benefit corporation doesn't sell: it asks for trust. I make mission, activities, and governance readable and citable, with an anchor to verifiable Wikidata references that protects from confusion and gets you found by those who can support you.
Projects I believe in, before collaborations
Not everything goes through commercial logic. I follow and support initiatives that reflect how I see the future: digital skills as a common good, sustainability of B Corps and the nonprofit sector, remote work in harmony with local territories. They are idea labs before they are partnerships, where technology is a tool for generating shared value.
A social innovation network promoting territorial regeneration and rural networking through nature-based co-working and remote work. Within the core team I handle digital infrastructure design, CRM integration, and semantic visibility strategy development.
A nonprofit association created to bridge the gap between education and the digital job market. As Project Lead I guide students in building real technology projects to support the third sector. It's where I took my first steps and started experimenting with technology's social impact.
A benefit corporation and B Corp focused on developing digital solutions for social and environmental innovation. I collaborate as an external consultant on information architecture definition and semantic data organisation, with the goal of demonstrating that digital can be ethical, transparent, and oriented towards the common good.
A higher technical education institute focused on post-secondary technology training. I teach Digital Marketing and deliver practical modules on AI and Advertising within the Higher Technician course in Artificial Intelligence & Business Management, transferring concrete skills in semantic data structuring.
What is a Semantic Entity Architect?
A Semantic Entity Architect designs the data layer that makes a person or organisation comprehensible to machines. The role doesn't write content and doesn't do SEO: it builds the entity — the structured, verifiable set of who you are, what you do, and what you're connected to — that AI systems use to describe you.
It's a role born at the intersection of knowledge architecture, structured data, and narrative strategy. The technical output (JSON-LD, schema.org, knowledge graphs) is the final deliverable; the core is deciding what to say and sourcing the evidence that proves it.
SEO, GEO, and AI readiness: what's the difference?
SEO optimises for ranking in traditional search engines. GEO — Generative Engine Optimization — optimises for being cited in responses generated by LLMs. AI readiness is the prerequisite for both: making your identity readable and verifiable by machines.
The three coexist but don't coincide. You can rank first on Google and be invisible to ChatGPT, or vice versa. My work operates at the deeper level: the informational infrastructure on which how AI agents read you depends, regardless of the channel.
Let's talk
If the problem I describe sounds familiar, book a 30-minute call. You don't need to arrive with a clear idea of what you need: the call is exactly for that.