Concept
Semantic Layer
A governed meaning layer that removes ambiguity across business data, real-world representations, digital twins, knowledge graphs, and AI systems.
Position in the Trusted Data Framework
Semantic Layer
Living Graph View
Semantic Layer in the concept network
Drag nodes to adjust the view, double-click a node to expand its neighbourhood, and right-click a node to hide it while preserving the exploration context.
Relationship Matrix
Semantic Layer as a network node
A semantic layer should sit on governed, quality-checked, reusable data rather than every application rebuilding meaning from raw source systems.
Ontology provides the formal meaning; the semantic layer turns that meaning into operational definitions and reusable access patterns.
The semantic layer prepares consistent entity, metric, and relationship definitions that can populate or query a graph.
A semantic layer needs technical metadata, lineage, quality signals, and ownership information to stay trustworthy.
Semantic APIs expose business meaning directly, rather than forcing applications to understand source-system structures.
AI applications need a governed layer of meaning to avoid inconsistent interpretation and brittle prompt-specific logic.
Concept Relationships
How Semantic Layer works with other concepts
A semantic layer should sit on governed, quality-checked, reusable data rather than every application rebuilding meaning from raw source systems.
An AI assistant uses a curated asset dataset with approved identifiers and quality signals before applying semantic definitions.
Which source data needs to be curated once before semantic definitions can be reused many times?
Ontology provides the formal meaning; the semantic layer turns that meaning into operational definitions and reusable access patterns.
The ontology defines Asset and Location; the semantic layer exposes trusted asset metrics and location hierarchies.
Which ontology concepts need to become reusable definitions for delivery teams?
The semantic layer prepares consistent entity, metric, and relationship definitions that can populate or query a graph.
A governed Customer definition feeds customer relationships, account ownership, and interaction events into a graph.
Which semantic definitions should be materialised into connected knowledge?
A semantic layer needs technical metadata, lineage, quality signals, and ownership information to stay trustworthy.
Column lineage and data quality scores show whether a metric definition is fit for AI-assisted decision support.
What metadata is required to prove that a semantic definition is reliable?
Semantic APIs expose business meaning directly, rather than forcing applications to understand source-system structures.
An AI assistant calls an Asset Health API instead of reconstructing asset health from raw operational tables.
Which business concepts should be available as APIs rather than buried in databases?
AI applications need a governed layer of meaning to avoid inconsistent interpretation and brittle prompt-specific logic.
An executive AI assistant uses approved revenue, risk, and asset definitions instead of inferring definitions from report fragments.
Where does AI currently rely on implicit or inconsistent business meaning?
Definition
A semantic layer translates fragmented data structures into shared meaning. It creates a consistent way to define metrics, entities, relationships, representations, and rules so humans, systems, and AI can interpret the same thing in the same way.
Semantic Layer improves AI not because the model becomes smarter, but because ambiguity is removed.
Why It Matters
Without a semantic layer, each team builds its own interpretation of the data. This creates conflicting metrics, duplicated logic, and brittle AI applications.
In most organisations, semantic layer work starts with business analytics: customers, revenue, KPIs, metrics, BI, SQL, dashboards, and LLM access over a warehouse. That is valuable, but it is not enough for AI-ready digital infrastructure.
The Trusted Data Framework extends the semantic layer from business meaning into reality meaning: how physical objects, places, assets, events, digital twins, simulations, and agents are represented and reasoned over.
Business Semantic Layer vs Reality Semantic Layer
| Business Semantic Layer | Reality Semantic Layer | | --- | --- | | Customer | Reality Mapping | | Revenue | Identity | | KPI | Representation | | Metrics | Spatial Objects | | BI | Digital Twins | | SQL | Multi-resolution | | Dashboards | Ontology | | LLM over warehouse | Knowledge Graph | | Business definitions | Simulation | | Analytical consistency | AI Agents |
The Business Semantic Layer removes ambiguity in business analysis.
The Reality Semantic Layer removes ambiguity between the physical world, digital representations, semantic models, knowledge graphs, simulations, and AI agents.
Role in the Trusted Data Framework
The semantic layer sits between Curated Data and Knowledge, but it also connects back to Reality, Representation, and Identity. It is where trusted data and digital representations become interpretable enough for reasoning, automation, simulation, and decision support.
This follows a core framework principle: curate once, reuse many. Business and operational systems produce data; the data platform prepares it; semantic, AI, and application layers consume it with governed meaning.
Practical Examples
- A single definition of customer, asset, supplier, incident, or risk.
- Governed metric definitions reused across dashboards, models, and AI assistants.
- Semantic mappings between operational systems and analytical platforms.
- A digital twin exposes the same real-world asset consistently across BIM, GIS, IoT, documents, and simulation.
- An AI agent reasons over a building, road, or pipeline using persistent identity and governed representations rather than source-system labels.
Case Evidence
GBlocks exposes its semantic layer as live,
queryable infrastructure (GET /semantic-model) rather than a static
document -- the same pattern this framework's Semantic Readiness
Accelerator looks for. It is deliberately a Reality Semantic Layer, not
a Business Semantic Layer: it extends meaning across physical objects
(Property, Flood Zone, Boundary), identity, and spatial relationships,
not BI metrics or KPI definitions. Self-assessed at Maturity Level 3
("priority concepts, mappings defined for selected use cases") against
this framework's own semantic layer maturity scale -- explicitly not yet
Level 4, since there is no formal versioning or change-control process
governing edits to the model beyond git history.
The model has since been extended once under real pressure -- four new
concepts and two new relationship types added as new sources arrived --
without restructuring any of the original concepts, a small but genuine
signal that the Level 3 scope was drawn correctly the first time rather
than by accident of a small initial dataset. A changelog field was
added directly inside the model to record what changed and why: a
lightweight, honest step toward the missing Level 4 governance, short
of a full versioning process.
Consulting Questions
- Which definitions are currently duplicated across reports and systems?
- Which data products need shared semantic contracts?
- How are metric definitions approved and changed?
- Where should semantics live: ontology, data catalog, BI layer, graph, or API?
- How will AI applications retrieve trusted meaning?
- Does the semantic layer only describe business metrics, or does it also describe real-world entities and representations?
- Where does AI currently confuse a representation for reality?
- Which physical or operational concepts need semantic grounding before agents can act safely?