# Novartis scales AI from demos to biomedical research How data platform investments enable AI in drug discovery and safety prediction - **Source:** [McKinsey Insights](https://www.mckinsey.com/insights) - **Published:** 2026-09-21 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/14140d7f-cdc7-4fef-ba33-007a411824ad ## From pilot to production Novartis has moved beyond AI proof-of-concepts to operational deployments in biomedical research. The pharmaceutical giant now uses AI for critical functions like safety prediction and generative chemistry. This shift reflects maturation of both technology and internal data infrastructure. Key to this transition was building robust data platforms to support AI at scale. ## Safety prediction applications One major application area is predicting drug safety profiles using AI models. This helps identify potential adverse effects earlier in the development process. By flagging safety concerns sooner, Novartis can reduce late-stage trial failures. The approach leverages historical trial data and molecular characteristics to forecast outcomes. ## Generative chemistry advances Novartis employs generative AI to explore novel chemical compounds for drug development. The system proposes molecular structures with desired therapeutic properties. This accelerates the discovery phase by rapidly generating and evaluating candidate molecules. The technology complements traditional medicinal chemistry approaches. ## Translation speed benefits A key outcome has been faster translation from discovery to clinical development. AI helps prioritize the most promising candidates, reducing wasted effort on less viable options. This compression of timelines could significantly impact the typical 10+ year drug development cycle. Early results show measurable improvements in research efficiency. ## Data platform foundation Underpinning these applications is Novartis's investment in unified data platforms. These systems aggregate and standardize research data across the organization. High-quality, accessible data is essential for training effective AI models. The platforms also enable collaboration between research teams working on different projects. ## Operational challenges Scaling AI required addressing several operational hurdles. These included data quality issues, model validation processes, and integration with existing research workflows. Novartis had to develop new governance frameworks for AI use in regulated environments. Change management was equally important as the technical implementation. ## Future roadmap Novartis plans to expand AI applications to additional research areas. Potential targets include clinical trial design optimization and biomarker discovery. The company is also exploring partnerships with academic institutions and tech firms. Continued platform enhancements will focus on supporting more complex use cases. ## Industry implications Novartis's progress demonstrates AI's growing role in pharmaceutical R&D. Success requires both technical capabilities and organizational adaptation. Other drugmakers face similar challenges in scaling beyond pilot projects. The experience offers lessons for regulated industries adopting advanced analytics. ## Sources covering this story - https://www.mckinsey.com/industries/life-sciences/our-insights/novartiss-christian-diehl-on-scaling-ai-beyond-the-demo — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._