Every DQ rule is derived from the Canonical Metadata Model — traceable to its source field definition, business rule, and approval decision.
NOT NULL constraints, required field validations, and mandatory relationship checks — derived from field-level metadata and business rules in your STTM.
Primary key uniqueness, business key deduplication, and composite key validation rules — generated from the target data model and mapping specifications.
Foreign key relationship checks, lookup table validations, and cross-table consistency rules — derived from the entity relationships in your metadata model.
Numeric range checks, date format validations, enumeration constraints, and pattern matching rules — generated from field data type and domain metadata.
The AI layer suggests additional DQ rules based on field names, data types, domain patterns, and historical data engineering knowledge — surfaced for human review.
All generated DQ rules pass through a human review workflow. Engineers validate, annotate, and approve before rules are deployed to the target environment.