End-to-end document pipeline: PDF/image ingestion → Claude vision OCR → structured data extraction with schema validation → routing to downstream system (CRM / ERP / billing / filing cabinet). Handles invoices, contracts, loan docs, medical forms, claim documents, tax filings. Includes the rejection queue for low-confidence extractions, human-in-the-loop approval UI, and the audit trail every regulated industry requires.
Discovered on Internal via @thetalkinghedge