Which records and relationships can a generated sufficiency dataset contain?

A generated dataset for this workflow is built around the case file. A release can include these record types and the links between them:

- **Cases**: an application, claim, order or engagement with a type, amount, jurisdiction and status history - **Requirement sets**: the policy items that apply to each case type, with conditions, validity rules and allowed substitutes - **Documents**: generated items with type, issuer, dates, extracted text and quality attributes such as missing pages or illegible sections - **Satisfaction links**: which document meets, partly meets or fails which requirement, and why - **Correspondence**: information requests and replies, with timestamps and what each one asked for - **Reference data**: thresholds, product rules and exception policies that change the required set

The relationships matter more than the volume of records. A requirement that links to no document is a gap. A document that links to two requirements may satisfy one and fail the other. A reply that arrives after a status change tests whether the agent updates its view of the file.

Every record is generated, not copied from a real business. MirrorFirm's positioning is that structure and behaviour are informed by direct observation of how organisations handle these decisions, but buyers should still inspect grounding and disclosure risk in a sample, and should not assume that any generated record is free of resemblance to real material.