SRE-style runbooks, cost guardrails, and continuous performance tuning.

Add lineage, catalog, RBAC, encryption-at-rest/in-transit, and monitoring.

Migrate tables & procedures (SQL Server / Oracle → Snowflake/Databricks), redesign models for ELT and dbt, implement real-time ingestion where needed.

Inventory data sources, measure data quality, map pain points (siloed sources, high legacy cost, slow queries).