At GSK, a British drugmaker, routing a drug-safety report to the correct specialist consumes significant time. Reports arrive via emails, paper mail, faxes, and data loggers. Each document must be read, prioritized, and directed appropriately. Repeated data entry and reconciliation is also common as information passes between systems. Lisa-Ann Chiriseri, Director of Drug Safety and Pharmacovigilance Applications at GSK, explained to Newsweek that the main issue lies with data management.
GSK aims to transform the movement of information in pharmacovigilance, which involves collecting and evaluating drug safety data. The objective is to allow specialists more time for critical decision-making. Chiriseri elaborated on these initiatives during the Seeking Value video series presented by Cognizant, alongside Mohammad Haque, the Global Head of Intelligent Enterprise Platforms at Cognizant. Newsweek’s Gabriel Snyder moderated the discussion.
Cognizant collaborates with GSK on automating workflows in drug safety. According to Haque, refining individual tasks doesn’t suffice if subsequent teams must reconstruct or re-enter the same data. The focus is on tracing each safety report from arrival through processing and evaluation to regulatory reporting and beyond. This preserves information while reducing manual preparation steps.
Chiriseri noted GSK’s established safety systems remain pivotal but acknowledged they need not dictate every workflow. The quest is for efficient information flow without overhauling underlying systems. Automation decisions start with problem identification and assessing technology’s reliability in solving them. Certain repetitive tasks can be automated even if skilled specialists are currently engaged.
The company adopts AI-based and conventional automation for classifying safety reports, setting priorities, and directing data to relevant experts. Structured data helps maintain context for reviews. Haque envisions an expanded role for specialized AI agents throughout pharmacovigilance processes.
He proposed an operating model that combines Cognizant’s services with an AI-based platform covering seven interconnected areas of drug safety. This spans from receiving reports to evaluating safety signals and regulatory preparations. Approximately 40 specialized agents may eventually handle full life-cycle activities, such as intake, coding, information checking, and evidence retrieval. Parallel task execution reduces dependencies between teams.
Human judgment remains crucial due to regulatory complexities. Haque distinguishes tasks governed by set rules from those needing expert interpretation. Conventional automation can manage predictable tasks, while AI assists in processing complex assessments. Oversight includes agents preparing work before qualified professionals act and agents executing workflows with employee-set parameters.
Chiriseri emphasized audit trails for traceability, enabling information handling and decision reconstruction. Validation and monitoring ensure systems operate within limits. Regulations offer a framework for responsibly introducing new capabilities, ensuring safeguards are built during development.
The companies reassess how they value technology investments. Haque noted traditional metrics, such as software-handled activities, fail to capture full operational improvements. He advocates measuring outcomes like reduced case processing time and faster safety concern resolution.
Chiriseri observed that GSK is moving from pilot projects to operational use, with encouraging preliminary results from the collaboration with Cognizant. Success indicators include employee adaptation to technology. GSK encourages workers to become technology-proficient and directly partake in process improvements.
Chiriseri advocates reusing existing information and technologies, avoiding continuous capability rebuilding. Connecting data across systems supports new workflows without complete application replacements. Haque advises similar companies to map processes and data before AI agent introduction. “Let the agents handle predictable tasks, make decisions visible, and ensure accountability by design,” he concluded.