Semantics and standardisation revisited

General-purpose EUBWs will make it easy to exchange all sorts of data between organisations - APIs without need for technical integration and contracts. WeBuild has established a Semantics group to get the full standardisation  - automation benefits - from the wallet rollout.

I had to ask ChatGPT for an update on semantics and standardisation.

Semantics and content standardisation are deeply intertwined — but not the same thing. Here’s how they relate and reinforce each other:

1. Semantics = shared meaning

Semantics deals with what things mean — the concepts and relationships between them.
It’s the vocabulary and logic layer:

• Defining what “Company,” “Director,” “IBAN,” or “Power of Attorney” means.
• Ensuring everyone (and every system) uses these terms consistently.
• Usually expressed via ontologies (OWL), controlled vocabularies (SKOS), or semantic data models (RDF, JSON-LD).

Without semantic alignment, every actor can publish “data,” but it won’t be understood in the same way by others — or by AI agents.

2. Content standardisation = shared structure and syntax

Content standardisation defines how the data is packaged and exchanged.
It’s the format and structure layer:

• How a verifiable credential, XML message, JSON schema, or PDF is structured.
• What fields are mandatory, optional, or conditional.
• Examples: ISO 20022, UBL, W3C VC Data Model, UN/CEFACT Core Components, etc.

Without content standardisation, even if we agree on meanings, we can’t automate exchange or verification at scale.

3. Their relationship — the two halves of interoperability

Think of them as meaning + form:

Together they make data portable, verifiable, and automatable.

4. Why this matters for EU Trust Infrastructure

In the EU’s trust stack (EUDI, EUBW, Data Spaces, AI agents), semantics ensures that a “Proof of Employment” issued in Finland can be understood in France, while content standardisation ensures it can be technically verified and ingested.
Without both:

• AI-agents can’t reason reliably.
• Wallets can’t exchange attestations cross-border.
• Data spaces can’t federate meaningfully.

5. Ideal architecture

1. SKOS vocabularies → define concepts.
2. OWL ontology → express relationships and inheritance.
3. SHACL / JSON Schema → constrain how instances (content) are expressed.
4. Standardised containers (VC, XML, etc.) → exchange and verify the content.

That’s the semantic-to-content bridge:
Vocabulary → Ontology → Shape → Payload

In one picture: https://www.linkedin.com/posts/bo-harald-4768b51_in-one-picture-activity-7389899650704707585-quoS?utm_source=share&utm_medium=member_desktop&rcm=ACoAAABPj1oB9_D7YNYACmHvY9HioUqpuULqZCo