Merck Improves Speed and Reduces Manual Effort to Assess Change Proposals Up to 70% on AWS

Merck's chemistry, manufacturing, and control scientists had to manually research across multiple repositories and documents to assess thousands of annual change proposals against regulations from over 100 global agencies. Merck built the Change Assessment Knowledge Engine (CAKE) on Amazon Neptune, combined with an NLP model trained on historical data, to identify affected product markets and relevant regulations and generate recommendations with hyperlinks to source material. The pilot was developed and implemented in less than 10 months, reducing change-assessment duration by up to 90% and cutting total manual effort per assessment by 30-70%. Merck's John Baker, Executive Director of Machine Learning and Analytics, is quoted praising AWS for delivering long-term customer value.

AWS is Amazon's cloud platform, used across pharma for genomics compute, SageMaker/Bedrock model training, and healthcare-and-life-sciences-specific services like Amazon HealthOmics.

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.