Answer
Isolate the single failing record and reproduce the same operation through the normal application flow before assuming the migration tool itself is at fault — a job's bulk error often has the same underlying cause a manual entry would hit, just harder to see. See design for retries and failures.