Over the past years, I have been deeply involved in the practical implementation of artificial intelligence in day-to-day pharmacovigilance operations. My work focuses on real-world use of AI for case intake processing, E2B R3 validation, and workflow optimization with a strong emphasis on compliance, human oversight, and measurable efficiency gains. Several of these approaches are already implemented in production environments.
Pharmacovigilance case processing remains one of the most resource‑intensive tasks in drug safety, especially for small and mid‑sized companies with limited personnel and growing global workloads. This presentation showcases how a small organization successfully transformed its case intake process by collaborating on an AI‑driven solution capable of generating E2B(R3) XML files directly from natural‑language source documents. The talk provides an honest, practical walkthrough of the journey—from market exploration and early system limitations to continuous improvement, breakthrough moments, and the eventual establishment of a functioning Proof‑of‑Concept. Emphasis is placed on the critical role of Human Quality Control to ensure regulatory compliance, especially in light of requirements such as those outlined in the EU AI Act. The presentation also highlights the unique challenges of validating evolving AI systems, the integration of AI into end‑to‑end PV workflows, and the implications for risk assessment in SaaS environments. Participants will gain insights into common pitfalls, realistic expectations, and the measurable benefits achieved, including reduced manual workload, increased processing speed, and rapid iteration cycles driven by real‑world feedback. This session is designed for PV professionals, quality managers, and digital transformation leaders seeking a concrete, experience‑based understanding of how AI can be safely and effectively introduced into pharmacovigilance operations.
- How a small pharma company successfully built an AI‑powered E2B(R3) case intake process, achieving automation traditionally reserved for large organizations.
- A real‑world journey from pilot chaos to working Proof‑of‑Concept, including obstacles, failures, breakthroughs, and lessons learned.
- Demonstrating how Human Quality Control + AI can dramatically boost accuracy while staying compliant with EU AI Act Article 14.
- Insights into risk assessment, validation challenges, and regulatory expectations when integrating AI into pharmacovigilance workflows
- Concrete results: time savings, reduced manual workload, rapid development cycles, and a scalable approach to future PV automation.
.webp)
%2C%20Otsuka.webp)
