Purpose

The Semantic Layer enriches your existing data model with metadata that adds clarity, meaning, and context to your datasets. This enriched metadata powers both human understanding (via the Data Hub) and AI (via Composable AI Agents).

Objective	Description
Enrich Data Model with Metadata	Adds descriptive and operational metadata to your Snowflake datasets, enabling more intuitive data exploration and governance.
Provide Context in Data Hub	Enables business users to understand datasets in natural language, surfacing context such as definitions, lineage, and usage directly in Simon’s Data Hub.
Empower Composable AI Agents	Provides structured metadata and relationships that AI Agents can leverage to generate higher-quality responses, predictions, and insights across marketing workflows.

By giving structure and meaning to your data, the Semantic Layer helps marketers and AI-driven systems “understand” your business concepts—bridging the gap between technical data models and real-world customer understanding.

Implementation

The Simon Semantic Layer builds upon and extends Snowflake’s Semantic Models.

While Snowflake provides the underlying framework for defining metrics, dimensions, and relationships, Simon’s layer adds marketing-specific enrichment and automation to make the data immediately useful across customer activation, analytics, and AI workflows.

Foundation: Snowflake Semantic Models

At its core, Simon’s Semantic Layer reads and extends your existing Snowflake Semantic Models.

These models define:

Entities (e.g., customers, orders, campaigns)
Relationships (e.g., customer has many orders)
Metrics (e.g., lifetime value, conversion rate)
Dimensions (e.g., region, acquisition channel)

Simon leverages this foundation to automatically map key concepts into its marketing data schema, ensuring consistency across segmentation, activation, and reporting.

📘
Learn more about Snowflake’s semantic model specifications here.

Enrichment and Data Profiling

Simon extends Snowflake’s native model with data profiling—an automated process that scans connected datasets to detect and classify metadata for each field. This includes:

Enrichment Type	Example
Field Classification	Identifies field types (e.g., email, customer_id, order_date)
Semantic Tagging	Tags fields with marketing concepts (e.g., “loyalty tier,” “churn risk,” “product category”)
Relationship Mapping* (Upcoming)	Automatically links tables using primary and foreign keys
Metric Discovery* (Upcoming)	Detects aggregatable fields (e.g., revenue, orders_count)

This profiling enables Simon to surface semantic insights directly in the UI and make the data actionable across the platform—whether you’re building a segment, configuring a flow, or generating an AI-driven campaign brief.

AI Context for Composable Agents

By exposing the Semantic Layer’s metadata to Simon’s Composable AI Agents, the platform can:

Understand your business metrics, KPIs, and data definitions
Generate context-aware SQL queries
Produce explainable reasoning about trends and customer behaviors
Recommend segment definitions, audience opportunities, and next best actions

This contextual intelligence dramatically improves the accuracy and utility of AI-driven features in Simon—reducing manual setup and ensuring every suggestion or output aligns with your company’s data definitions.

Sharing and Collaboration

Semantic views can also be shared directly via Snowflake’s secure data sharing capabilities. This enables collaboration between data and marketing teams and allows you to distribute enriched, business-ready datasets across your ecosystem.

📘
For more on sharing semantic views, see Snowflake’s guide here.

Summary

Feature	Description	Benefit
Metadata Enrichment	Extends Snowflake models with marketing context	Improves data understanding and governance
Profiling and Discovery	Automatically identifies entities, metrics, and relationships	Reduces manual modeling and speeds onboarding
AI Enablement	Provides structured data context to AI Agents	Enhances AI accuracy and marketing insights
Data Sharing	Publishes enriched views through Snowflake	Encourages cross-team collaboration

Next Steps

Confirm your deployment supports the Semantic Layer (Composable Data Hub required).
Work with your Simon account manager to identify candidate datasets and define enrichment scope.
Review opportunities to share semantic views with partner systems or internal analytics teams.

📘
For guidance on enriching your data model and enabling semantic metadata in Simon, contact your account manager.