Amgen

Amgen describes using historical trial and real-world data to build a digital-twin synthetic control for each enrolled patient in AI-enabled trials.

Amgen's ATOMIC tool ranks candidate trial sites with machine learning, and Amgen reports that top-ranked sites enrolled patients markedly faster.

At its Breda plant Amgen uses cameras, analytics, chatbots over maintenance reports and ML disruption prediction to run operations more efficiently.

Amgen process development staff use generative AI tools over a curated technical knowledge base of more than a million documents, with evaluations and risk-based validation.

Amgen aims for predictive models and digital twins across its manufacturing network; process-parameter monitoring had been shown only in proofs of concept at the time.

Amgen adopted OpenAI's ChatGPT Enterprise early and widened access across the company in phases, with security controls.

Partners: OpenAI

Amgen's knowledge management team built a Q&A agent in Copilot Studio that searches the company's large body of R&D documents and reports.

Partners: Microsoft

Amgen was an early Microsoft Copilot adopter and expanded access from a small pilot group to about 20,000 employees.

Partners: Microsoft

Amgen installed an NVIDIA supercomputer called Freyja at its deCODE genetics subsidiary to train AI models on its very large store of de-identified human data for research.

Partners: NVIDIA