Beyond the dashboard: an MCP-connected AI agent that reads your dbt lineage

BI copilots can answer questions about what's on the screen: filters, dimensions, chart values. But ask "Why is this number wrong?" or "Where does this column come from?" and they hit a wall. They can't see past the dashboard.

We built an AI agent that goes deeper. Directly within the BI layer, it connects to your dbt project through MCP. The agent extracts context from the report, reads compiled SQL, traces column-level lineage from mart to source, explains the business logic in each transformation, and runs validation queries against the warehouse to confirm what the data shows.

For analysts, it eliminates hours spent following DAGs and reading model code. For business users, a persona switch delivers plain-English answers with no dbt jargon in sight.

We'll show how the dbt MCP server makes this possible and demo the agent live by investigating a data discrepancy from BI report to root cause to Jira ticket, all in one chat.