Snowflake Horizon Context
Active context for AI and BI that gives you trusted answers. Capture business semantics across BI and data systems so every AI agent, tool and application operates from the same trusted logic.
why now
End the metric debate
When “Active Customer” means something different in the data warehouse, the BI tool and the AI assistant, teams stop trusting the numbers. AI trained on raw tables without governed business definitions is far more likely to misinterpret what the data means, producing answers that look credible but use the wrong logic. And governance controls built around where data lives stop applying the moment logic moves to another platform. Horizon Context is designed to solve these issues.
Collect context
Pull metadata from disparate systems to give AI a complete picture of data inside and outside Snowflake.
Enrich context
Automate metadata enrichment and reduce manual effort to build and maintain a context layer while keeping humans in the loop.
Activate context
Discover, access and automatically activate data context for AI, BI and apps.
Key Differentiators
One source of truth across every tool and use case
Connectors and standards
Collect metadata across many sources
Gather rich context, such as query logs, popularity and schemas from many BI sources into one searchable catalog.
Go beyond Snowflake-only metadata. Catalog your full data estate with out of the box connectors, such as Tableau, Power BI* and more.
Configure any OpenLineage producer (like Apache Airflow1) to send lineage information automatically.
Share metadata across vendors with the Open Semantic Interchange (OSI) standard, backed by 50+ participating companies.
Business context layer
Enrich your data to make every asset AI-ready
Surface quality signals from column-level lineage across Snowflake and external databases, and auto-generate documentation from metadata.
Build semantic views in Semantic Studio,* an AI-assisted IDE in Workspaces with Git-based versioning and Snowflake CoCo integration.
Skip manual work with Semantic View Autopilot, ingest existing SQL, Tableau or Power BI files,* and generate your semantic view automatically.
Handle level-of-detail metrics, composable definitions and materializations that rewrite queries for you.
Multi-platform reach
Surface trusted answers in every tool your teams use
- Find assets across your entire data estate with Universal Search,* now faster with hybrid keyword and semantic ranking.
Query governed definitions natively from Power BI, Tableau, Excel, Google Sheets, Google Data Studio and ThoughtSpot.*
Ask any data question and get answers grounded in governed definitions. CoCo finds and queries the right semantic views automatically.
Give external AI agents governed data access through MCP. Connect from Claude, Cursor or any agent framework.†
Powered by Snowflake Horizon Catalog
AI agents and conversational analytics
Ask any data question in natural language and get answers grounded in governed business logic. Snowflake CoCo and Snowflake CoWork automatically discover and query the right semantic views, so every AI agent and analyst works from the same trusted definitions.
Semantics
Define your business logic once in Semantic Views and it applies everywhere. Use Semantic Studio* to build and test definitions with AI assistance, or let Semantic View Autopilot generate semantic views automatically from your existing SQL, Tableau or Power BI files.
Out-of-the-box BI connectors for rich metadata
Catalog your entire data estate — from Snowflake to BI tools like Tableau and Power BI — into a single searchable hub enriched with query logs, schemas and popularity metrics.
Trusted answers with Snowflake CoCo
CoCo automatically retrieves relevant context using hybrid keyword and semantic search.
“Horizon Context will give us the foundation to deploy AI agents that automatically inherit our evolving business logic, keeping every system aligned without manual intervention.”
Brendan Cyrus
Director of Product, AI Analytics & Data Platforms, Indeed
"Snowflake’s Horizon Context helps extend consistent business definitions across our broader data ecosystem, supporting more trusted and governed AI and analytics experiences at scale.”
Dan Shah
Manager of Data Science
Horizon Context Ecosystem Partners
HORIZON CONTEXT
Frequently Asked Questions
Find answers about setup, compatibility and how Horizon Context fits your existing stack.
Horizon Context is the governed context layer built into Snowflake Horizon Catalog. It centralizes business definitions, metrics, relationships and logic, so they apply consistently across every BI tool, AI agent and application.
Yes. Pull context from external databases, BI tools and pipeline systems using metadata connectors. Governance policies apply regardless of where data physically resides.
Horizon Context improves accuracy by allowing AI agents to automatically discover and query semantic views that contain governed business logic. Instead of guessing context from raw schemas, AI models reason from trusted definitions, reducing hallucinations and producing consistent, reliable answers.
No. dbt defines transformation logic (how data is built). Semantic views define business meaning (how data is interpreted). They are complementary. Use dbt to transform; use Horizon Context to govern meaning.
Third-party layers sit outside the query engine. Governance can be bypassed by querying tables directly. Horizon Context enforces security and business logic at the engine level — it cannot be circumvented.
The context Snowflake gathers spans several layers. You can think about it this way:
- Structural: what exists and how it connects (ex. tables, columns, lineage)
- Operational: what’s happening (ex. queries, freshness, performance)
- Semantic: what it means (ex. definitions, metrics, ontology)
- Behavioral: how it’s used (ex. popularity, query patterns)
