dbt model documentation and test drafts
Generate column descriptions, tests and semantic definitions for dbt models, then have the model owner review before merging.
DifficultyBeginner
Setup time1-2 hours
CategoryData & analytics
The use case
dbt projects often have models with no descriptions and few tests because documenting is tedious. dbt's AI assistant in the dbt platform, now called dbt Wizard, can generate documentation, data tests, semantic models and SQL from the model code and project context. A data engineer reviews descriptions for accuracy, especially business meaning that is not visible in code, and checks that tests reflect real expectations rather than current data. Changes go through a pull request like any other dbt change.
Example workflow
- Pick undocumented or untested models
- Generate descriptions and tests
- Fix business meaning the code cannot show
- Run dbt build to check tests
- Merge through a pull request
Tools
- dbt
- dbt Wizard
Risks and guardrails
- Descriptions can sound right but misstate business meaning
- Tests based on current data can bless bad data
- Review in PRs like hand-written changes
Sources
Written by Agents Lookup. Credits and official pages for the tools this use case names:
- dbt docs: dbt Wizard: docs.getdbt.com/docs/platform/wizard-platform
- dbt MCP server (official page): github.com/dbt-labs/dbt-mcp (checked 2026-11-09)