Novo Nordisk Accelerates Clinical Insight with Custom Agents on Azure

Despite harmonizing over 200,000 patient-years of clinical trial data through its FounData initiative, Novo Nordisk still needed months of coordination among specialists and biostatisticians to extract insights, slowing early-stage decision-making and increasing the risk of pursuing unviable hypotheses. Working with Microsoft's Forward Deployed Engineering team, Novo Nordisk built a governed reasoning agent on Azure that lets researchers explore clinical hypotheses independently: the agent generates and executes code against proprietary datasets with AI-generated statistical analysis, explainable outputs, and human-in-the-loop validation by biostatisticians before insights inform major decisions. This cut time to insight from weeks to minutes for many exploratory analyses, and the team can now evaluate 50+ strong ideas per quarter versus 5-10 previously, reducing low-value exploratory work while maintaining scientific rigor; expansion is underway toward predictive capabilities and broader data integration.

Microsoft Azure is Microsoft's cloud AI platform, combining Azure Machine Learning and Azure OpenAI Service — the delivery channel for a large share of pharma's enterprise GPT deployments.

This sits in the Build stage of the AI use-case lifecycle: Named examples from the foundation-model labs and platforms pharma teams use to actually construct, train and generate the model or application.