Three small trials. Real checks.

What these examples establish

Each scenario has a fixed source and two target snapshots. The corrected trial uses the same rules. The check never silently fixes identifiers or chooses the first duplicated row. A changed rule or source digest means a new baseline is needed.

These synthetic examples demonstrate behavior. They do not prove compatibility with every CRM / ERP export, endorse production cutover, or contain client data. The product is limited to company, contact and product master data.