Snowflake Intelligence is more than a data platform—it’s a layered architecture that transforms raw data into actionable insight. This conceptual model illustrates how databases, logical tables, semantic views, and orchestration agents work together to support verified queries and intelligent search. By mapping each entity, we gain clarity on how intelligence is structured and delivered.
| Entity Name | Description |
|---|---|
| Databases | Structured repositories that store raw and processed data. |
| Schema | Organized definitions of tables and relationships within a database. |
| Tables | Collections of rows and columns representing structured data. |
| Columns | Individual fields within a table that define data attributes. |
| Logical Tables | Abstracted views of data used for modeling and analysis. |
| Logical Table Columns | Sub-types of Logical Tables including Dimensions, Time Dimensions, Facts, and Metrics. |
| Relationship | Defined connections between logical tables for query resolution. |
| Cortex Analysis Models | Semantic layers enriched with instructions for intelligent analysis. |
| Verified Queries | Pre-approved query patterns that ensure reliable and secure data access. |
| Cortex Search Services | Search interfaces that interpret queries and return relevant results. |
| Cortex Search Service Columns | Sub-types including Authorization metadata for secure access control. |
| Agents | Components that manage orchestration, instruction, and user interaction. |
| Example Questions | Representative queries used to guide and test system responses. |
| Tools | Functional modules such as Cortex Analysis, Custom Tools, and Search Filters. |
| Search Results Filter | Mechanisms for refining and prioritizing returned data based on relevance. |
Understanding the architecture behind Snowflake Intelligence reveals how data becomes insight. Each layer—from schema to semantic view—plays a role in orchestrating clarity, speed, and trust.
Comments
Post a Comment