Drug development remains one of the most resource-intensive and failure-prone processes in healthcare. Traditional approaches depend heavily on trial-and-error experimentation across long timelines, and the integration of artificial intelligence (AI), including large language models and generative AI, is beginning to reshape how targets are identified, candidates are designed, and trials are executed.
This webinar, hosted by Charlotte Guzzo, COO of Sano, brought together leaders working at the intersection of AI and precision medicine to examine where these technologies are delivering measurable change and where significant challenges remain. In this blog, we explore key takeaways from the conversation.
Key Takeaways
- Accelerated Discovery: Generative AI is significantly reducing drug development timelines by optimizing target identification and candidate design.
- Data Challenges: Data scarcity and quality, particularly in rare diseases, remain primary hurdles for training robust AI models.
- Lab Automation: The future of biotech relies on integrating AI with automated laboratory synthesis to speed up testing cycles.
- Precision Medicine: AI is enabling better matching of complex biological pathways with treatments for chronic and rare diseases.
- Cross-Industry Innovation: Leadership talent is shifting from big tech to healthcare to apply computational expertise to medical breakthroughs.
About the speakers
Charlotte Guzzo
Charlotte Guzzo leads operations at Sano Genetics, with a focus on patient experience in precision medicine research, digital trial workflows, and at-home genetic testing. Her work focuses on enhancing patient experience in medical research through digital solutions and at-home testing. Before joining Sano, she researched childhood cancer origins at the Wellcome Sanger Institute and worked in risk management at JPMorgan Chase.
Parker Moss
Parker Moss leads corporate development at Exscientia, focusing on AI-driven drug design in oncology. His past roles include Chief Partnerships Officer at Genomics England and leadership positions in healthcare technology. His work at Exscientia supports AI-driven development of targeted medicines.
Margo Georgiadis
Margo Georgiadis is the CEO of Montai Health, where she leverages AI to create medicines for chronic diseases. A former leader at Ancestry and Google, Margo is known for her role in tech and genomics transformation. Her work at Montai Health centres on applying AI to the identification of biological targets for chronic disease, drawing on large-scale genomic and phenotypic datasets.
Guillermo Del Angel
Guillermo Del Angel, the Executive Director and Rare Disease Therapy Area Head at AstraZeneca's Center for Genomics Research, utilises his expertise in computational biology and data science to improve genomics-based treatments. With experience at Alexion Pharmaceuticals and the Broad Institute, he focuses on integrating advanced data analysis and genomics into patient care.
Highlights from the webinar
Personal journeys to the intersection of technology and healthcare
The webinar emphasized how personal experiences are driving the shift toward AI-powered precision medicine. Key transitions include:
- Parker Moss: Transitioned from telecommunications to oncology-focused precision medicine following his daughter's cancer diagnosis.
- Margo Georgiadis: Moved from leadership roles at Google and Ancestry to healthcare, focusing on AI for clinical trials and chronic disease treatments.
Challenges and innovations in AI-driven drug discovery
AI adoption in drug development is accelerating. The FDA has noted a significant increase in drug application submissions that include AI components, spanning nonclinical, clinical, postmarketing, and manufacturing phases. Against that backdrop, the panel examined where AI integration is already delivering results and where structural barriers remain.
Challenges discussed include:
- Data scarcity and quality: A significant challenge highlighted was the scarcity and quality of data needed to train AI models effectively. For rare diseases, as mentioned by Guillermo, the data can be particularly sparse, making it difficult to train robust models. This scarcity extends to chemical data as well, where the scale of explorable chemical space far exceeds what is captured in available chemical libraries.
- Integration with existing systems: The integration of AI tools with traditional drug development processes poses significant logistical and technical challenges. For instance, transitioning from AI models to practical applications requires not just technological solutions but also changes in regulatory, clinical, and operational frameworks.
- Computational demands: The computational power required to process large datasets and run complex simulations is enormous. As Parker discussed, generative AI and other advanced AI methods demand substantial computational resources, which can be a limiting factor in their deployment.
Innovations discussed include:
- Generative AI for drug targeting and design: Parker Moss described the use of generative AI at Exscientia to optimize drug targets and candidates more efficiently, which may reduce the time and cost of drug development by enabling earlier-stage screening and filtering of candidate compounds. This technology allows for pre-screening of drug compounds, ensuring that only the most promising candidates are synthesized and tested.
- AI in precision matching of chemistry and biology: Margo discussed how Montai Health uses AI to match complex biological pathways with potential treatments. This approach is particularly relevant for chronic diseases, where understanding and modulating biological pathways may enable more targeted treatment options for chronic diseases.
- Automation and efficiency in laboratories: A major innovation is the automation of laboratory processes. As Parker pointed out, the future of drug discovery involves automating synthesis and testing in the lab alongside computational modeling. Automating these processes reduces manual variability, shortens iteration cycles, and produces more reproducible results across discovery stages
- Advanced screening techniques: AI-driven methods are enabling more sophisticated screening techniques that can handle vast chemical libraries and biological datasets. These techniques can predict drug efficacy and safety profiles earlier in the drug development process, potentially reducing the high rates of failure in later stages.
- Machine learning in genomic data interpretation: Guillermo highlighted the use of machine learning to interpret genomic data, particularly in rare diseases. This application of AI helps in faster and more accurate diagnosis and patient identification, which is crucial for effective treatment development.
- AI in clinical trial design and post-market monitoring: Beyond discovery, AI is increasingly applied to later stages of the drug development workflow. Recent reviews have mapped AI's role across the full pipeline, including preclinical and clinical study optimization and post-market surveillance. For precision medicine programs, where patient identification and eligibility confirmation are complex, these applications are particularly relevant to reducing cycle times and improving data quality.
Summary
The webinar examined where AI is delivering tangible progress in drug development and where structural barriers, particularly around data quality, regulatory integration, and computational scale, continue to limit adoption. Speakers brought perspectives from oncology, rare disease, chronic disease, and genomics, illustrating that the most meaningful advances come from aligning AI capabilities with specific biological and operational problems rather than applying them broadly.
For sponsors running precision medicine programs, the implications are direct. AI is reshaping how targets are identified, how candidates are designed, and how patients are matched to therapies. As regulatory frameworks evolve to accommodate AI-generated evidence, the ability to integrate these tools into trial design, recruitment, and data workflows will become a practical requirement rather than an advantage.