AstraZeneca

AstraZeneca authors describe Research Assistant, an internal LLM tool for exploring biomedical questions, and share lessons from running it at scale in R&D.

AstraZeneca built a supervisor-led multi-agent Brand Assistant on Databricks that gives commercial teams governed access to structured and unstructured data, growing from a small proof of concept to more than 20 agents in production.

Partners: Databricks

AstraZeneca runs a company-wide AI upskilling programme, and employees who reach a set accreditation level get early access to an internal AI agent builder, which produced over 1,500 agents in two months.

AstraZeneca, alongside Astellas and BMS, agreed to use and share operational data with Evinova's AI-native trial platform (an AstraZeneca health-tech business) to get benchmarks and recommendations that speed trials.

Partners: Evinova

AstraZeneca built Development Assistant on AWS, a multi-agent tool that lets development teams ask plain-language questions of their data and get answers quickly.

Partners: AWS

AstraZeneca bought Modella AI, following a 2025 collaboration, to bring its pathology foundation models and AI agents in-house for oncology R&D, including biomarker and targeted-therapy work.

Partners: Modella AI

AstraZeneca agreed with Tempus and Pathos to build a large multimodal oncology foundation model from clinical and molecular data to support its cancer R&D.

Partners: Tempus, Pathos

AstraZeneca and Pangaea Data will co-develop a multimodal AI platform, using Microsoft and NVIDIA technology, that finds undiagnosed or under-treated patients and surfaces therapy and trial options inside clinical workflows.

Partners: Pangaea Data, Microsoft, NVIDIA

AstraZeneca built a knowledge graph combining public adverse-event and exposure data to speed COVID-19 vaccine safety analyses and periodic reporting.