Create interpreter score records for program coordinators before billing

Your AI grading runs arrive as fragmented LLM responses causing missing scores and delayed certification records. Coordinators and billing get completed score and narrative records in your database before the next billing run.