AI database query context brief - editable editorial worksheet
This is a downloadable guide, not an automatic product import or built-in template.
Never put credentials, secrets, or unapproved row data in this document.

Task and decision:
Database engine and version:
Approved tables/views and execution identity:
Entity being counted (row grain):
Business term and exact definition:
Relevant fields, types, keys and relationship cardinality:
Status/event codes included and excluded:
Date field, reporting timezone, inclusive start and exclusive end:
Units and currency (if applicable):
Null/unknown and empty-result treatment:
Grouping, sort order, tie handling and row limit:
Fields to retain during follow-up questions:
Question plus independently reviewed SQL example:
Reference result and how it was checked:
Ambiguity requiring clarification:
Unsupported dimensions or unavailable data:
Definition owner, version and review date:

Training context is not an authorization boundary or a guarantee of correctness.
Use the product's current documentation and SQL example fields where supported.
Separate training examples from held-out evaluation questions.