The hidden catalyst for AI breakthroughs in drug discovery
06 May 2026
Data Quality
- One of the issue with training models is having access to enough data of sufficient quality and well annotated. What solutions do we have for that?
- AI has changed how we generate molecules and hit identifications is not the hardest part in drug discovery any more. How will AI support the next steps, experiments, work with experimental data? How will data need to be structured? For AI to create break through time and cost savings, how will it be applied in wet labs, instruments? When AI works with experimental data how will this change data and privacy considerations? Will and how will AI impact and morph current (experimental)workflows?
- Where does a system of record fit into the agentic workflow loop and is this a commonly overlooked aspect?
- Are proprietary datasets still a competitive advantage, or is that advantage eroding?
- What “hidden catalyst” today will become obvious in hindsight five years from now?
