How Ecolab used an AI migration agent to shrink 84,000 lines of legacy SQL by 6x
Blog post from dbt
Ecolab is modernizing a highly complex legacy data platform containing more than 400 interdependent facts and dimensions sourced from over 3,000 systems, where Snowflake tasks, JavaScript stored procedures, and incremental processing had created opaque logic and unresolved data-quality issues. Working with Datafold, the company used an AI migration agent to map lineage, incorporate institutional knowledge from internal materials, translate legacy transformations into over 40 dbt projects, and validate each rebuilt model through automated row-level data comparisons. For a flagship dimension, the effort reduced 409 models to 170, eliminated about 70,000 lines of code and 1,000 data connections, while the overall migration has produced roughly six times less code and four times fewer connections. Git-based version control, automated testing, and deployment also made the resulting workflow closer to SOC 2 readiness, while visual data diffs helped gain stakeholder approval for around 90 tables in the first four weeks. Ecolab found that the migration required an iterative approach, with structural requests emerging after teams saw working dbt projects, and plans to retain its current reporting layer until all dimensions and facts have been reviewed before switching underlying tables.
No tracked trend matches for this post yet.
Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.