An interesting discussion on AI in drug discovery and development with Thirupathi Pattipaka from...

An interesting discussion on AI in drug discovery and development with Thirupathi Pattipaka from Novartis, Kimberly Noel M.D. M.P.H from Roche and Meriem Sefta from OWKIN at HLTHEurope2025.

One example from Novartis is using agents for the generation of real world evidence, significantly cutting down the time for this process. They also use AI for patient identification and patent population selection in their clinical trial design, and in the preparation of clinical trial dossiers.

Owkin discussed how to leverage AI for patient stratification and the development of dedicated biomarkers. In one example AZ did a retrospective analysis of a failed trial showing that an AI enabled biomarker may have turned that trial into a success and is now redoing this trial with the new biomarker. They also highlighted that the cost of AI implementation will go down over time as economies of scale come in.

The panel also discussed the still emerging regulatory aspects of the new AI approaches, as well as the important topics of data bias in the underlying data, particularly for any algorithms touching patient data.

That said there are plenty of opportunities to leverage AI, and particular LLMs, to increase operational efficiency of the process, keeping a human in the loop. This is an ongoing process that will require some significant change management with the pharma workforce.