DenmarkHealthcare and life sciences2026 – ongoingNovo Nordisk
Azure reasoning agent for clinical-data analysis
Novo Nordisk's clinical-analysis agent reduced exploratory analysis from weeks to minutes, Microsoft reported
Problem
Novo Nordisk had harmonized clinical datasets, but answering research questions still depended on scarce statistical expertise. Slow analysis delayed decisions about which hypotheses deserved further work.
Approach
Novo Nordisk worked with Microsoft's forward deployed engineers to build a reasoning agent on Azure. Researchers asked questions in natural language, and the agent generated and executed analytical code against proprietary clinical data. Biostatisticians and scientific experts validated promising outputs before major decisions.
Results
Takeaway
Prepared clinical datasets supported the faster analysis. Scientists retained responsibility for interpreting the outputs.
Categories
- Size
- Large enterprise
- Industry
- Healthcare and life sciences
- Business function
- IT and R&D
- Innovation type
- Process innovation
- AI technology
- Generative AIAI agents
- Adoption
- Partner co-development
- Evidence
- Vendor case study
Sources
Source summaries
Microsoft described a deployed clinical-analysis agent with human validation and automated evaluations. It reported exploratory analyses taking minutes instead of weeks.
Full details are in the original sources.
