Effectiveness of generative AI comes down to data and prompting.
A common pitfall to implementing generative AI in writing contexts is a misunderstanding of how LLMs interpret prompts and analyze the data exposed to it. Source data is the lifeblood of regulatory submissions, but the complexity of data and file formats (docx, PDF, tables, figures) can impact the quality of AI generated content for your submission.
This session will provide best practices for leveraging generative AI against various source data files. This will include a close-look at various data sources used to develop documents within the submission dossier, how each data source can be leveraged differently by generative AI, and how users can best prompt the generative AI model to receive desired results.
Key Learning Objectives:
-
- Understand the nuances of prompt engineering for the effective use of generative AI.
-
- Define how best to use generative AI based on source data available and desired outputs.
-
- Identify where and when generative AI is effective in the submission drafting process.
Speakers: Nick Brown and Liam O’Leary