Which records can be generated for a conditional approval dataset?

A conditional approval dataset is a set of linked records, not a single table. A release can include:

- **Decision records** with the approver role, the scope of what was approved, the date and the policy that applied. - **Condition records** with rule text, owner role, due point, whether the condition blocks a step, and a status history. - **Evidence records** such as certificates, reports, forms and emails, each linked to the condition it is meant to satisfy. - **Review records** showing who assessed the evidence and what they found, including rejections and referrals. - **Change records** for conditions that are varied, waived, added or lapse. - **Downstream requests**, such as a drawdown, a work order or a purchase release, that depend on the conditions. - **Policy and role material**, including who may waive what and which conditions are mandatory.

The relationships carry the value. A condition belongs to one decision, evidence may fit more than one condition, and a waiver may be valid only if the right role authorised it. A task pack can then ask questions that need those links followed.

A simulated environment adds state on top: a queue of cases at different stages, an inbox where evidence arrives over time, and tools that change record status.

All of this is generated. It is informed by direct observation of how businesses run these handovers, but a buyer should still inspect a sample for grounding and disclosure risk.