Novartis: Accelerating Drug Development with AI-Powered Clinical Trial Transformation
NovartisAmazon Web Services (AWS) · Databricks · Amazon S3 +8
Novartis partnered with AWS Professional Services and Accenture to modernize their drug development infrastructure and integrate AI across clinical trials, with the goal of reducing trial development cycles by at least six months. The initiative built a GXP-compliant data mesh platform on AWS with Databricks for processing, enabling AI use cases including protocol generation and an intelligent decision system (digital twin). Early results from the patient safety domain showed 72% query speed improvements, 60% storage cost reduction, and 160+ hours of manual work eliminated. The protocol generation use case achieved 83-87% acceleration in producing compliant protocols.
Overview
Novartis partnered with AWS Professional Services and Accenture to modernize their drug development infrastructure and integrate AI across clinical trials, with the goal of reducing trial development cycles by at least six months. The initiative built a GXP-compliant data mesh platform on AWS with Databricks for processing, enabling AI use cases including protocol generation and an intelligent decision system (digital twin). Early results from the patient safety domain showed 72% query speed improvements, 60% storage cost reduction, and 160+ hours of manual work eliminated. The protocol generation use case achieved 83-87% acceleration in producing compliant protocols.
The challenge
Novartis's drug development and clinical trial data was fragmented across heterogeneous sources — file shares, relational databases, life sciences platforms like Veeva Vault, and master data management systems like Reltio — hindering AI adoption across the R&D continuum in an industry where developing new medicines typically takes 15 years. Novartis set the goal of reducing clinical trial development cycles by at least six months.
The solution
Novartis partnered with AWS Professional Services and Accenture to build a next-generation GXP-compliant data platform on AWS with Databricks for processing, following a data mesh architecture with five components: an ingestion framework grouping similar source types, a three-tier data product storage/processing model, data management and governance (catalog, lineage, quality), self-service data consumption via BI tools and AI/ML platforms, and a central observability platform. The platform is designed to support AI use cases including protocol generation and an Intelligent Decision System (digital twin) for simulating clinical trial operational plans; compliance was built in from the design phase using Architectural Decision Records and threat modeling rather than added afterward.
Reported business value
The patient safety domain, an early adopter, built 16 data pipelines processing approximately 17 terabytes of data, yielding a 72% reduction in query speeds, a 60% reduction in storage costs, and 160+ hours of manual work eliminated. The protocol generation use case, which currently runs on demo data or legacy on-premises systems, achieved 83-87% acceleration in producing protocols that meet compliance standards. The six-month clinical trial reduction goal remains to be fully proven, but early results suggest substantive progress toward that target.
Sources
Open any source and check the claim yourself — that is the point of the register.
This record was researched and written with AI assistance, and its claims were checked against the sources above. (EU AI Act art. 50 transparency notice.)
Other life sciences entries in the register.
Novo Nordisk accelerates clinical documentation and drug development with Claude
Novo Nordisk, a global pharmaceutical company headquartered in Europe, built NovoScribe, an AI-powered documentation platform using Claude Code, Amazon Bedrock and MongoDB Atlas, to automate clinical study report (CSR) generation. Claude helped cut writing times on CSRs by 90%, reducing time spent producing clinical study documentation from 10+ weeks to 10 minutes, and delivered a 95% reduction in resources needed to create device verification protocols. The platform expanded from clinical trial reports to device protocol documentation and patient materials, and an 11-person development team, including non-technical staff, now prototypes features using Claude Code.
Novo NordiskClaude Code · Amazon Bedrock · MongoDB AtlasProtecting Crops Today to Feed Billions Tomorrow
Syngenta built Gaia, the first crop protection R&D data platform for analytics and innovation, on Databricks to unify siloed data across 70-plus countries. The platform cut data access times from as long as nine months to minutes, reduced time to value by 70%, decreased engineering costs by 50%, and now publishes over 200 data products handling more than 2 million queries per month.
SyngentaDatabricks · Delta Lake · Unity Catalog +3Novo Nordisk uses machine learning and computer vision to automate pharmaceutical manufacturing quality checks
Novo Nordisk, a Danish pharmaceutical company that supplies nearly 50 percent of the world's insulin, built an automated ML pipeline on AWS using Amazon SageMaker Pipelines to train, deploy, and monitor computer vision models on edge devices. The system automates previously manual tasks including counting drug cartridges in boxes and detecting bacterial anomalies on agar plates, using a robotic arm, camera rig, and edge inference, with results monitored via an Amazon QuickSight dashboard. The solution deploys ML models at scale to different edge devices, automates quality-assurance tasks, improves time to market, and was successfully repurposed from the cartridge-counting use case to the agar-plate anomaly detection use case with minimal changes.
Genome Institute of Singapore accelerates whole genome processing with Databricks
Genome Institute of Singapore (A*STAR), working with Precision Health Research Singapore (PRECISE) on the National Precision Medicine strategy, used the Databricks Data + AI Platform with Delta Lake and Glow to run joint calling of 10,000 whole genomes in under 72 hours, a 15x improvement over their prior approach, to support population-scale precision medicine research.
Was this helpful?
Your feedback helps us improve our use case database