Pfizer PACT Initiative: 14 AI Projects Saving 16,000 Research Hours in Pharmaceutical Development

Title

Pfizer PACT Initiative: 14 AI Projects Saving 16,000 Research Hours in Pharmaceutical Development

Content

With a mission to transform healthcare for its 1.3 billion patients, Pfizer launched the Pfizer-Amazon Collaboration Team (PACT) initiative with AWS in 2021 to drive innovation in digital drug development. PACT has pursued 14 projects including generative AI and machine learning to save scientists up to 16,000 hours of search time annually and cut infrastructure costs by 55 percent. Pfizer uses a variety of AWS tools to explore new ideas, quickly build prototypes, and drive innovation in drug development. The first use case minimized time spent in data discovery by up to 80 percent for 1,500 pharmaceutical sciences scientists. The development of one drug can result in approximately 20,000 documents, and scientists previously had to search for data manually using various tools. Early experimentation led to the development of Amazon Kendra, an intelligent enterprise search service. PACT teams now use generative AI, accessing Anthropic's Claude 2.1 through Amazon Bedrock via an internal platform called Vox for voice-activated document search and natural language querying. Another PACT prototype detects anomalies in Pfizer's Portable Continuous Miniature and Modular (PCMM) continuous manufacturing process for producing solid oral-dose medicines. The PCMM anomaly detection prototype uses Amazon SageMaker for ML capabilities. Pfizer scientists can detect anomalies as they occur, predict maintenance needs, and reduce equipment downtime using Amazon Lookout for Equipment and Amazon Lookout for Metrics. This work led to a manufacturing analytics tool specializing in time series data that has been adopted by Pfizer's large molecule group. PACT has cut the time from prototype to MVP from 3 months to no more than 6 weeks, and from the initial 14 projects, five ideas have moved into production.