Agentic AI APQR in pharma

LG CNS’s Agentic AI APQR in pharma provides the most convenient and efficient service for generating APQR reports. It significantly reduces the time required for data collection and processing—from what used to take several months using manual or semi-automated methods—allowing customers to focus their investments on enhancing the quality of the reports themselves. 

LG CNS’s Agentic AI APQR in pharma is a service that manages the entire process—from collecting raw data scattered across various systems such as EDMS, LIMS, ERP, and QMS, as well as handwritten paper documents, to preprocessing and final report generation. To support APQR generation, agents are structured around the planning, execution, and verification phases. Each execution unit corresponds to the detailed business process segments defined within the PQR.

This service leverages a managed infrastructure based on Amazon EKS to easily deploy and reliably operate multiple Agent services. Generative AI-based Agents are implemented using Amazon Bedrock, which enables integrated management of models and knowledge bases, while automating the full process from execution planning to operation and evaluation. Model responses are cached using Amazon ElastiCache, improving response times and enabling cost optimization. Context data between Agents is shared via Amazon S3 or Amazon RDS, ensuring session continuity and data consistency. For data collection and processing, Amazon Athena provides a fast and scalable data pipeline. In addition, high-accuracy search functionality is implemented through Retrieval-Augmented Generation (RAG) using OpenSearch, which is effectively applied to the management of GMP and SOP documents — critical compliance requirements in the pharmaceutical industry. User authentication is securely and scalably managed through Amazon Cognito, enhancing both the security and user experience of the overall application.