Create parsed ID lists for data scientists before refresh

Your webhook payloads arrive with compressed ID lists, causing missed records and delayed model updates. Data scientists receive clean ID lists and counts for downstream pipelines so feature updates run before the next model refresh.