Beyond the Prompt: Exploring AI-Assisted PK/PD Workflows with the Monolix Agent
  • 8:00 AM Pacific / 11:00 AM Eastern
  • Online
  • Register

Abstract:

Just because a task can be delegated to an AI agent doesn’t mean it should be.

In pharmacometrics, the same task can often be done directly in the software, automated with a script, or delegated to an AI agent. Each approach involves different trade-offs in flexibility, reproducibility, and the effort required from a modeler. And as AI capabilities evolve, these choices will change.

In this webinar, we will walk you through a PK/PD modeling and simulation example using the Monolix agent in Composer. From data preparation and model development to covariate analysis and simulation, we will move between the agent, the graphical interface, and scripted automation, exploring when and why each approach makes sense.

We will also examine what is required when an agent works on your behalf: what the modeler needs to be able to inspect and control, and how systematic testing can help assess and maintain agent reliability as the underlying large language models, tools and software change.

As AI agents become increasingly integrated into pharmacometric workflows, scientists will face new choices about how to perform and oversee their analyses. This webinar offers a practical starting point for making those choices deliberately, rather than simply adapting to whatever the technology makes possible.

Presenter:

Monika Twarogowska – Sr. Principal Scientific Product Specialist