Introducing dbt Charts: Dashboards as code

Your data pipelines are automated. Your transformations are versioned, tested, and reviewed. Why should the dashboard be any different?

AI can generate a chart. The harder problem is giving AI the context to understand your data and produce answers your business can trust.

dbt Charts brings dashboards into the development lifecycle with an open-source, declarative language for BI. Board files live next to your dbt models, can be authored by humans or AI agents, reviewed in a pull request, and validated before they reach your users. And because dashboards are defined as code, AI has a machine-readable layer for understanding how your data should be queried, visualized, and shared.

In this session, we’ll take a dashboard from nothing to shipped. You’ll see how SQL and YAML work together, how AI can build and modify dashboards using trusted data context, and what happens when the underlying data or models change.

See how dashboards as code can help your organization get more from its data and give humans and AI a shared, governed foundation for understanding it.

What you’ll walk away with:

• How to give AI better context around your company’s data, creating a shared, governed foundation that helps both people and AI get more value from it.
• How humans and AI agents can build together using SQL and declarative YAML to create, modify, and ship dashboards.
• Why dashboards as code matter in the AI era and why AI needs more than access to data to deliver trustworthy answers.
• How dbt Charts extends analytics engineering practices through the BI layer, bringing version control, testing, review, and CI to dashboards.

Meet the speakers

RJ Andrews

Principal Chartographer

dbt Labs

David Fowler

GM, dbt Charts

dbt Labs