Validation & data quality

Validation agents enforce data quality via configurable rules. Rules can be simple required-field checks (title present, at least one image, at least one variant) or AI-powered semantic rules (e.g. 'description is at least two sentences, mentions the product, and contains no placeholder text', or 'the primary image actually shows the product').

Rule-based validation consumes zero credits; only AI semantic rules consume credits. Validation results surface in the Review Queue and on the Command Center as items needing attention. SKUuz seeds a sensible default rule set for new organisations.