Select Star attacks a problem every maturing data team recognizes: the data exists, but nobody is sure which version of it is authoritative or who to ask. The platform connects to your warehouses, ELT/ETL pipelines, and BI tools, reads their metadata and query history, and automatically produces documentation, lineage, and usage statistics. Instead of a manually maintained spreadsheet of tables, you get a living catalog that reflects what is actually being queried, including which tables have quietly gone unused.
Lineage is the centerpiece. Select Star traces data down to the column level and across platforms, so you can answer concrete questions: if this field changes, which dashboards break, and who gets affected? The same graph is used for governance — PII tags can be propagated to downstream columns so sensitive data is visible wherever it flows, which matters for audits and compliance work. A data portal layer sits on top, letting anyone search for certified datasets, read descriptions and tags, and record discussions against a dataset so answers stay attached to the data.
More recently Select Star has leaned into AI readiness. Reporting is that it is now part of Snowflake, and the product exposes metadata, lineage, and generated semantic models to LLMs and agents through interfaces such as an MCP server, aiming to be the context layer that makes AI answers trustworthy rather than plausible. Pricing is quote-based rather than published, and is typically sized against the number of connected sources, data volume, and required seats, so evaluation starts with a sales conversation.
