Publish
Publish your completed Semantic Model to make it available for business use. ekai supports publishing to Snowflake Intelligence (Cortex Agents), data catalogs, or downloading as a complete artifact package.
Complete the Execute DBT step to build and run your DBT models.
Publish & Review Tab
Navigate to the PUBLISH & REVIEW tab to see publishing options:
Publishing Options
Snowflake Intelligence
Status: ✅ Available
Publish to Cortex Agents for natural language querying of your data.
- Cortex Analyst integration
- Conversational AI agents
- Auto semantic view creation
Data Catalogs
Status: On-Demand
Publish to OpenMetadata or custom catalog integrations.
- OpenMetadata (standard)
- Custom integrations (enterprise)
- Contact sales for setup
Artifact Download
Status: ✅ Available
Download complete package for local use, manual Git, or archival.
- DBT project
- All documentation
- Lineage diagrams
Snowflake Intelligence Publishing
Configuration
Configuration Fields
| Field | Description | Example |
|---|---|---|
| Assistant Name | Identifier for the Cortex Agent | V_Baltic_Bridge_Bank_Customer_360_View_ |
| Description | What the agent does | "Snowflake Cortex Agent for Customer 360" |
| Access Control | Who can use the agent | Everyone or specific roles |
| Select Roles | Snowflake roles (if specific) | ACCOUNTADMIN, ANALYST |
| Compute Warehouse | Warehouse for queries | COMPUTE_WH |
| LLM Model | AI model for the agent | CLAUDE-SONNET-4-5 |
| Semantic View Name | Auto-generated (read-only) | V_Baltic_Bridge_Bank_Customer_360_View_ |
Access Control Options
| Option | Description |
|---|---|
| Everyone in my organization | All Snowflake users can access |
| Only specific roles | Limited to selected Snowflake roles |
LLM Selection
You can override these settings using the LLM Model dropdown.
Publish
Click Publish To Snowflake Intelligence to deploy.
After Publishing
Success State
Your assistant is now available for Q&A within the Snowflake Intelligence interface.
Post-Publish Actions
Data Catalog Publishing
OpenMetadata Integration
For data catalog publishing:
| Field | Description |
|---|---|
| OpenMetadata Server URL | e.g., http://localhost:8585 |
| Security Token | JWT token for authentication |
Click Connect & Publish to deploy catalog metadata.
What Gets Published
- Technical metadata (tables, columns, types)
- Business glossary terms
- Data lineage information
- Metrics and KPI definitions
For custom catalog integrations, contact sales@ekai.ai.
Artifact Download
Download a complete package containing all generated artifacts:
Package Contents
| File | Format | Description |
|---|---|---|
| Data Description | .md | Overview documentation |
| BRD | .md | Business Requirements Document |
| DBT Project | .md + files | Complete dbt project |
| ERD | .dbml | Entity Relationship Diagram |
| Lineage | .yaml | Data lineage definition |
| Profile | .md | Statistical profile summary |
Use Cases
| Use Case | Description |
|---|---|
| Local Development | Run dbt locally |
| Manual Git | Commit to repository |
| Sharing | Share with external teams |
| Backup | Archive for compliance |
| CI/CD | Integrate with pipelines |
Re-Publishing
To update a published model:
- Make changes (BRD, artifacts)
- Re-run DBT if needed
- Return to PUBLISH & REVIEW
- Click Re-Configure & Publish
The existing Snowflake Intelligence agent is updated in place.
Summary
| Destination | Availability | Use Case |
|---|---|---|
| Snowflake Intelligence | ✅ Available | Natural language querying |
| OpenMetadata | On-demand | Data governance |
| Artifact Download | ✅ Available | Development, backup |
Next Steps
- Security & Privacy — Data protection details
- Schema Agents — Create more ERDs
- Create another Model — Build more data products