返回 Agenda
Case Study Sponsored by Clinevotech: Show Me the Evidence: Build Inspection-Ready AI in Pharmacovigilance with Proven Controls, Validation & Real Use Cases
Session Chair(s)
Invited Speaker
, Switzerland
As AI moves from pilots into GxP pharmacovigilance, inspectors no longer ask whether you use AI; they ask you to show the source, model version, confidence and human reviewer behind every AI-assisted decision. This session translates the latest expectations from CIOMS Working Group XIV, FDA, EMA and the ISPE GAMP® AI Guide into practical controls for hallucination, LLM drift, reproducibility, source traceability, AI validation (CSV), human-in-the-loop oversight, explainability, confidence scoring and model versioning. Drawing on real implementations across call and mail intake, literature monitoring, signal management and Health Canada regulatory case intake for tier-1 and tier-2 pharma companies and CROs, the speaker will share what worked, the pitfalls to avoid and the efficiency gains achieved without compromising compliance. Attendees will leave with a reference architecture, an inspection-readiness checklist and the key questions every MAH should ask GenAI PV vendors.
Learning Objective : - Apply a risk-based framework to place AI and set human oversight in PV
- Control hallucination, drift and irreproducibility with source citations, pinned models and regression testing
- Trace any AI-populated ICSR field to its source, model version and reviewer
- Build a lifecycle CSV strategy aligned with GAMP 5 and the ISPE GAMP® AI Guide
- Ask GenAI PV vendors the right questions
- Learn from real use cases delivering 70%+ efficiency gains
Speaker(s)
Ashish Bhandarkar, MPharm
Vice President – PV Product Development, Client Engagements & Lead SME-PV, Clinevo Technologies Pvt. Ltd., India