Retail Semantic Ontology Layer¶

The Semantic Ontology Layer sits between the Data Plane and the Application Catalogue, providing a unified, standards-aligned knowledge graph that resolves entities across POS/Commerce, CRM/CDP, ERP/Supply Chain, E-commerce/Digital Analytics, Social Media/Reviews, and In-Store IoT systems.

Industry Standards Alignment¶

The retail ontology is aligned to global product identification and retail industry standards.

GS1 Standards¶

The GS1 system provides globally unique identifiers for products, locations, and shipments. Every product, store, and logistics entity in this ontology carries GS1-aligned identifiers:

GS1 Standard	Purpose	Ontology Entity
GTIN	Product/SKU identification	`Product_Master.GTIN`
GLN	Store/warehouse location	`Store_Master.GLN`, `Warehouse.GLN`
SSCC	Shipment tracking	`Shipment.SSCC`
GPC	Product classification	`Product_Master.GPC_Code`

NRF / Retail Taxonomy¶

The National Retail Federation (NRF) retail taxonomy and Global Product Classification (GPC) govern process areas and product hierarchies:

Process Area	Scope	Ontology Coverage
Merchandising	Assortment, pricing, promotions	Demand Forecasting, Price Optimization, Supplier & Merchandising
Store Operations	Labor, planogram, shrinkage, energy	Store Operations
Supply Chain	Inventory, fulfillment, procurement	Inventory & Fulfillment, Demand Forecasting
Customer	Loyalty, personalization, engagement	Customer 360, Personalization
Digital Commerce	E-commerce, mobile, marketplace	Personalization, Inventory & Fulfillment
Sustainability	Waste, carbon, traceability, compliance	Sustainability & Compliance

Architecture¶

graph TD
    subgraph DataPlane [Data Plane - 6 Source Systems]
        POS["POS / Commerce"]
        CRM["CRM / CDP"]
        ERP["ERP / Supply Chain"]
        ECOM["E-commerce / Digital Analytics"]
        SOCIAL["Social Media / Reviews"]
        IOT["In-Store IoT"]
    end

    subgraph OntologyLayer [Semantic Ontology Layer]
        SCHEMA["Schema<br/><i>12 Domains · ~50 Tables</i>"]
        WF["Workflows<br/><i>10 Retail-aligned</i>"]
        POL["Policies<br/><i>8 Governance Rules</i>"]
        INT["Integrations<br/><i>6 System Mappings</i>"]
    end

    subgraph Apps [Application Catalogue - 8 Apps]
        C360["Customer 360"]
        DEMAND["Demand Forecasting"]
        PERSON["Personalization"]
        PRICE["Price Optimization"]
        STORE["Store Operations"]
        INV["Inventory & Fulfillment"]
        SUPPLIER["Supplier & Merchandising"]
        SUSTAIN["Sustainability & Compliance"]
    end

    POS --> SCHEMA
    CRM --> SCHEMA
    ERP --> SCHEMA
    ECOM --> SCHEMA
    SOCIAL --> SCHEMA
    IOT --> SCHEMA

    SCHEMA --> WF
    POL --> WF
    INT --> SCHEMA

    WF --> C360
    WF --> DEMAND
    WF --> PERSON
    WF --> PRICE
    WF --> STORE
    WF --> INV
    WF --> SUPPLIER
    WF --> SUSTAIN

Ontology Components¶

The ontology consists of four interconnected layers, all stored in enterprise-knowledge/:

Layer	Files	Format	Purpose
Schema	`retail-schema.yaml`	YAML	12 domains, ~50 tables with fields, types, risk levels, constraints, and relationships
Workflows	`workflows/*.yaml` (10 files)	YAML	Triggered process flows with steps, rules, SLAs, and dependencies
Policies	`policies/*.md` (8 files)	Markdown	Business rules, thresholds, approval chains, and compliance constraints
Integrations	`integrations/*.yaml` (6 files)	YAML	Field-level sync mappings between source systems with conflict resolution

Relationship Types¶

Entities in the schema are connected via typed relationships that form a knowledge graph:

Relationship	Meaning	Example
`triggers`	Entity event initiates a workflow	`Transaction` quantity spike triggers `Demand Reforecast` workflow
`syncs_to`	Entity data flows to another system	`Product_Master` syncs_to `E-commerce_Catalog` via integration
`constrained_by`	Entity operations governed by a policy	`Promotion` constrained_by `pricing-policy`
`depends_on`	Entity requires a parent/related entity	`Inventory_Position` depends_on `Store_Master`
`validates`	Entity validates another in a business process	`Vendor_Master` validates `Product_Master` for sourcing

Cross-Application Entity Heatmap¶

Shows how many of the 8 retail apps require each entity. Higher usage = more foundational.

Entity	Count	Applications
`Product_Master`	8	All apps
`Transaction`	6	Customer 360, Demand Forecasting, Personalization, Price Optimization, Store Operations, Revenue
`Store_Master`	5	Demand Forecasting, Store Operations, Inventory & Fulfillment, Price Optimization, Sustainability & Compliance
`Customer_Profile`	4	Customer 360, Personalization, Price Optimization, Revenue
`Inventory_Position`	4	Demand Forecasting, Inventory & Fulfillment, Store Operations, Sustainability & Compliance
`Promotion`	3	Demand Forecasting, Price Optimization, Personalization
`Vendor_Master`	2	Supplier & Merchandising, Sustainability & Compliance
`Shelf_Condition`	1	Store Operations

Product_Master is the most connected entity in the retail ontology — equivalent to Material_Master in Supply Chain, Customer_Master in Banking, Patient_Master in Healthcare.

Schema Statistics¶

Metric	Count
Domains	12
Tables	~50
Source Systems	6 (POS/Commerce, CRM/CDP, ERP/Supply Chain, E-commerce/Digital Analytics, Social Media/Reviews, In-Store IoT)
Workflows	10
Policies	8
Integrations	6
Relationship Types	5 (triggers, syncs_to, constrained_by, depends_on, validates)
GS1 Standards Covered	4 (GTIN, GLN, SSCC, GPC)

Documentation¶

Document	Description
Schema Reference	Complete domain and table reference with fields, types, and relationships
Workflows	All 10 AI workflow definitions with triggers, steps, and dependencies
Policies	All 8 governance policies with rules, thresholds, and approval chains
Integrations	All 6 system integration mappings with field-level detail
App-Object Mapping	Minimum required objects per application with relationship matrix
Mapping Architecture	Schema-first architecture, Semantic RAG pipeline, and ReAct tools

← Back to Retail Catalogue