Case Study: Cognizant achieves faster, more accurate purchase order processing and secure PII/PHI redaction with UiPath Document Understanding and AI Center

How Healthcare Organizations Unlock Productivity by Processing Semi-Structured Documents with UiPath Document Understanding and AI Center

Cognizant worked with a large New York–area healthcare provider that was overwhelmed by semi-structured and scanned documents—especially purchase orders and supplier credit memos—and needed to securely redact PII/PHI before sharing claims. Manual data entry averaged about eight minutes per document, suppliers sent data in many inconsistent formats, and credit memo backlogs and compliance risk were rising. To address this, Cognizant deployed UiPath solutions including AI Center and Document Understanding integrated with its RPA engine.

Using UiPath, Cognizant built automated pipelines for credit adjustments and document redaction, repurposing an invoice ML model for credit memos and retraining it with ~250 examples from 20% of suppliers that accounted for ~80% of incoming data. Results included faster, more accurate processing, reduced backlogs, and progressively less human intervention; model training time fell by 40% (from 18 to 10 hours) after moving from CPU to GPU. UiPath’s Document Understanding, Taxonomy Manager and Action Center enabled reliable extraction, secure PII/PHI redaction, and human-in-the-loop validation to drive continuous improvement.

Cognizant

Shaun Dawson

AVP, Global Leader of IPA Center of Excellence