Enter Note Done
Go to previous page in this tab
Event
  • Scientific Session
  • Bookmark Icon
  • Transforming Drug Discovery Using Agentic AI for Target Knowledge Reviews and Weight of Evidence Decision-Making
  • Schedule
    Notes
  • Stars at Night Ballroom B1‚ Convention Center
    Set Timezone


Chair: Cody Robarts, Pfizer Inc.
Co-Chair: Zhichao Liu, Boehringer Ingelheim

Primary Endorser: Computational Toxicology Specialty Section


Endorser(s): Drug Discovery Toxicology Specialty Section


The rapid evolution of artificial intelligence (AI) in toxicology has shifted from static predictive models toward adaptive, multi-step agentic workflows capable of automating complex scientific reasoning tasks. These agentic systems, composed of coordinated AI agents that plan, retrieve, analyze, and synthesize information, offer the potential to fundamentally transform knowledge generation, regulatory documentation, and decision-making processes across drug discovery and safety assessment.

However, despite growing interest, the practical implementation of agentic AI in toxicology remains nascent, with key challenges related to transparency, validation, reproducibility, and integration into existing scientific workflows. At the same time, emerging use cases such as Target Knowledge Reviews (TKRs) and weight of evidence (WOE) assessment present ideal opportunities to evaluate the strengths and limitations of agentic approaches in real-world toxicology applications.

This session will consist of three 20-minute presentations followed by a 15-minute Q&A session. These presentations will explore how agentic workflows are being applied to accelerate and standardize knowledge synthesis and evidence integration in toxicology. The session begins with a foundational introduction to agentic AI systems, including their architecture, capabilities, and limitations in scientific contexts. Building on this foundation, the second presentation will illustrate how agentic workflows can be deployed in TKRs to systematically extract, structure, and summarize complex biological and safety data, improving both efficiency and traceability. The final presentation will examine how agent-driven frameworks can support WOE assessments in drug discovery by integrating diverse data sources into structured, defensible safety assessments.

Together, these presentations provide a cohesive narrative from conceptual foundations to applied case studies and demonstrate how agentic AI can enhance reproducibility, speed, and scientific rigor in toxicological decision-making. Attendees will gain practical insights into both the use cases and limitations of agentic systems, as well as considerations for responsible adoption in regulated environments.



IAT ITS Designation: Innovations in Applied Toxicology (IAT)
Session and Activity Type
Presentations