From DML soup to dbt structure: Ecolab's migration story

Ecolab's legacy data pipelines were a constant headache: tangled, unknown lineage meant small changes took weeks to implement and test. Silent errors like uncaptured source deletes crept into production, and stream-based processing made failures painful to recover from. The team wanted to migrate to dbt for clear lineage, readable code, and idempotent runs, but that meant translating over 2,000 DML statements by hand. To tackle this migration, Ecolab partnered with Datafold, whose AI process converts natural language requirements into automated checks, translates legacy logic into dbt, and validates the result against the original pipeline before sign-off. Migrations don't have to choose between fast and trustworthy. In this session, you'll see how to achieve both.