What linked records could a generated payment follow-up dataset contain?

A generated dataset can contain more than an invoice marked overdue. It can represent the relationships needed to establish which obligation remains unpaid and which account is responsible for it. Useful records include invoices, credit notes, receipts, payment allocations, account hierarchies and authorised billing contacts.

The relationships carry much of the evaluation value. One receipt may cover several invoices, while a credit note applies to only one line. A parent account may fund a subsidiary's invoices without being the approved recipient of collection messages. Treating either relationship as interchangeable can produce an incorrect demand.

MirrorFirm could generate a linked record graph in which amounts reconcile across these objects. A task pack could ask an agent to identify the outstanding obligation, explain the balance calculation and select the contact associated with that obligation.

The buyer should specify which relationship types are in scope and which ambiguities are intentional. An unexplained broken link is a dataset defect; an explicitly constructed missing allocation can be a valid investigation task. Keeping that distinction visible helps evaluation teams assess reasoning about incomplete records without accidentally rewarding guesses caused by inconsistent generation.