How should a reseller expose complaint event order to downstream systems?

A generated complaint history can use an append-only event representation. Useful fields include `event_id`, `case_id`, `event_type`, `occurred_at`, `recorded_at`, `actor_role`, `previous_state` and `next_state`. The difference between occurrence and recording time matters when evidence reaches the handling team after the underlying event.

For an insurance administration complaint, a document may have arrived before an acknowledgement but been indexed afterwards. A downstream system should reason from what was accessible at the decision point, rather than reconstructing an unrealistically complete view from the final file.

A reseller can distribute events as JSONL with a linked case index, or as Parquet tables for analytical use. Preserve identifiers and ordering rules across both representations. Do not rely on file position alone to communicate event sequence.

MirrorFirm could generate these structures with an accompanying event dictionary. The reseller's catalogue can state whether buyers receive occurrence order, knowledge order or both. This gives customers a concrete basis for choosing sequence modelling, workflow replay or historical decision evaluation, and prevents format conversion from silently removing the process information that gives the material value.