Data products
Identifies the right tables, checks the SQL and registers the result as a data product in the data catalogue. For data teams who need a governed product, not a one-off query.
Describe the product and see each step before anything is registered.
- Azure SQL databases (live, read-only)
- Collibra data catalogue
- A validated query for the data product
- A delivery ticket
- A catalogue entry for the product
Four steps from start to finish.
- 1Describe
A steward asks for a data product in plain language.
- 2Find
The agent identifies the relevant tables in the connected Azure SQL databases.
- 3Validate
It runs the query read-only. Only a single SELECT is allowed, and writes are blocked.
- 4Register
It opens a Jira ticket and registers the product in Collibra. The steward sees each step and can redirect it.
A data product starts as a query in someone's notebook and never gets registered, owned or documented.
A steward describes the product they need, for example a sales performance view. The agent finds the relevant tables, validates a read-only query, opens a delivery ticket and registers the product in the catalogue.
What it means for a team like yours.
Build a sales performance product from the retail tables. Illustrated with fictional demo data, not client data.
Register an invoicing view across invoices and sales orders. Illustrated with fictional demo data, not client data.
Combine trade tables that share a deal identifier into one governed view. Illustrated with fictional demo data, not client data.