Objectives - Lay the necessary foundations to activate Snowflake Intelligence:
Structured ingestion,
Bronze → Silver → Gold model,
Semantic layer
Cortex Analyst configuration
Exposure via Cortex Agent
Snowflake Intelligence configuration
Testing
Knowledge transfer
Deliverables - Study report including:
High-level target architecture
Ingestion pipelines
Bronze, silver, and gold data models
Data normalization in the silver layer
Semantic layer and Cortex Analyst created
Cortex Agent was created based on Snowflake Intelligence
Prerequisites:
Access to necessary data
Access to your Snowflake account or use of a trial account
Example questions to be answered by the agent
Availability of a contact person for 1 hour/day for 5 days
Assumptions:
Source data is understandable, or a data model is provided
No more than 10 source tables containing only structured data
Customers concerned:
Organisations already on Snowflake but still in the structuring phase, wishing to quickly activate self-service agentics uses.
Commercial proposal:
Flat fee of €5,500
Data Engineer + Snowflake expert
Approach | Activities |
|---|---|
Day 1 – Framing & Architecture | Quick analysis of the Snowflake stack (data organization, RBAC, warehouses, sources). Definition of the data scope for Snowflake Intelligence usage. Target architecture: ingestion → models → semantic layer → Cortex Analyst → Cortex Agent → Snowflake Intelligence. Validation of access rules and security constraints (access policies, masking policies). |
Day 2 – Ingestion & Bronze Model | Implementation of ingestion pipelines (batch loading, Snowpipe). Minimal data normalisation (format, conventions, typing). Documentation of the schema and quality control points. Knowledge transfer. |
Day 3 – Silver & Gold + Data Quality | Usage-oriented Gold modeling by Cortex Analyst. Knowledge transfer |
Day 4 – Semantic Layer + Cortex Analyst | Definition of the semantic view for the gold layer: metrics, dimensions, relationships, and business labels. Creation of the Cortex Analyst service and integration with a Cortex Agent. Configuration of FinOps limits (e.g., alert thresholds) and best practices for interpretation. Knowledge transfer. |
Day 5 –Snowflake Intelligence | Exposure of Cortex Agent in Snowflake Intelligence. UI configuration (logo, theme, welcome messages). End-to-end testing: NL queries → reasoning → SQL → visualization. Knowledge transfer. |
Pimp My Snowflake Semantic Layer in 2 days
Objectives:
Create or modernise the Snowflake Semantic Layer to expose a Cortex Agent using tools such as Cortex Analyst for structured data and Cortex Search for unstructured data.
Deploy this agent at the Snowflake Intelligence level to provide access to already structured BI data.
Deliverables:
Study report including: Diagnosis of the existing gold layer
Identification of topics that can be activated in natural language
Creation of the semantic layer, cortex analysis, and cortex agent
Configured Snowflake Intelligence interface
Knowledge transfer
Prerequisites:
Study report including: Diagnosis of the existing gold layer
Identification of topics that can be activated in natural language
Creation of the semantic layer, cortex analysis, and the cortex agent
Configured Snowflake Intelligence interface
Knowledge transfer
Assumptions:
Gold data is documented (column comments or access to a data catalog)
No more than 10 gold tables containing only structured data
Target customers:
Organisations that already have a structured warehouse (Silver/Gold tables, BI dashboards) but no formal semantic layer.
Commercial proposal:
Flat fee of €2,350 Data Engineer + Snowflake expert
Approach | Activities |
|---|---|
Step 1 – Diagnosis of the existing analytical layer | Review of existing models (SQL files, views, marts). Analysis of KPIs, dimensions, relationships, and naming conventions. Data quality verification. Identification of priority use cases. |
Step 2 – Design & deployment of the Semantic Layer | Construction of the semantic view (measures, dimensions, relationships, business labels). Compliance with Cortex Analyst guidelines (limits, query verification, aliases). Implementation of tests and validations (verified queries). |
Step 3 - Cortex Agent & Exposure in Snowflake Intelligence | Creation of the Cortex Analyst service. Implementation of Cortex Agent. Publication in Snowflake Intelligence: UI configuration, access rules, and user testing. Knowledge transfer. |