Shrink the surface: governed self-service for analyst-owned dbt pipelines

Most self-service dbt programs fail by giving analysts a full project and hoping governance catches up. You get unreviewed SQL in production and engineers debugging code they didn't write. We used dbt mesh to give analysts a separate AI-native project, isolated from our production codebase. Analysts own business logic and metric grain; the platform owns the rest, microbatch config, schema isolation, CI, downstream registration. 10+ analysts have shipped production pipelines without touching deployment or backfill logic. AI makes this boundary work. I built Claude skills, AI agents encoding dbt, microbatch, and git knowledge into repeatable steps. Analysts bring SQL; skills scaffold models, generate tests, manage PRs, and enforce conventions. This talk covers how we drew that boundary, the failures that reshaped it, and what I'd change.