Intelligent label processing with AI and LLMs improves compliance

A leading MedTech equipment manufacturer partnered with Sigmoid to automate the extraction and validation of UDI attributes from product labels for compliance with evolving global regulations.

Business Scenario

A global medical devices manufacturer needed to improve compliance with evolving global UDI (Unique Device Identification) regulations, especially the European Union’s EUDAMED requirements. The mandate to populate over 130 structured UDI attributes per product by 2027 posed a significant challenge, given the unstructured nature of product label data. Manually extracting and validating this data was not only labour-intensive but also error-prone, risking compliance, audit delays, and operational inefficiencies.

Sigmoid Solution

Sigmoid developed an AI-powered label analysis solution that automates the extraction and validation of UDI label attributes. The solution integrated computer vision, ML, and LLMs with a human-in-the-loop mechanism for accurately extracting and processing variable label formats. The solution was built to handle scale, accuracy, and future extensibility across business units. A multi-stage pipeline ingested label data from both SAP Everest and Tahiti systems, followed by advanced image pre-processing and AI-powered OCR for identification.