AstraZeneca Develops LLM-Based Research Assistant for R&D.

Piotr Grabowski, Mohamed Alameen, Jorge Bretones, Sabina Cardell, Miguel Carmona, Gavin Edwards, Ben Grainger, Sameh Hassan, Erik Jansson, Artur Kuziakhmetov, Albert Maristany, Hebatallah Mohamed, Andriy Nikolov, Sebastian Nilsson, Mark O'Donoghue, James Pacileo, Ashiq Sultan, Alex Voegele, Michael Ughetto· August 14, 2026 View original

Key takeaways

  • AstraZeneca deployed an LLM-based system for R&D, integrating diverse data.
  • The system supports both direct Q&A and multi-step research tasks.
  • Responses are grounded in evidence and linked to original sources.
  • It enhances daily R&D workflows for scientists and clinicians.

Who benefits

PharmaceuticalsBiotechnologyHealthcareLife SciencesChemical

Summary

AstraZeneca has deployed "Research Assistant," an internal LLM-based system designed to help scientists and clinicians explore biomedical questions across diverse data sources. It offers a chat interface, grounding responses in retrieved evidence and linking back to original sources.

AstraZeneca has developed and deployed an internal LLM-based system named "Research Assistant" to support its scientists and clinicians in R&D. This agentic system provides a chat-style interface, enabling users to query and explore complex biomedical questions by integrating information from a wide array of sources. These sources include scientific literature, knowledge graphs, chemistry databases, clinical trials, safety resources, expression data, and internal experimental systems. The Research Assistant supports both a fast mode for direct question answering and a multi-step mode for more intricate research tasks. A key feature is its ability to ground responses in retrieved evidence, providing direct links back to the original sources. This allows users to verify information and delve deeper into the underlying data, enhancing trustworthiness and utility in daily R&D workflows.

Why it matters

This demonstrates a practical, enterprise-scale application of LLMs to accelerate complex research and development, offering a blueprint for other organizations in data-intensive fields.

How to implement this in your domain

  1. 1Identify internal knowledge silos: Pinpoint areas where researchers struggle to access consolidated information across disparate systems.
  2. 2Pilot an internal LLM-based search tool: Start with a limited scope, integrating a few key data sources for a specific research team.
  3. 3Design a user-friendly chat interface: Ensure the system is intuitive and supports both quick queries and multi-step investigations.
  4. 4Prioritize evidence grounding and source linking: Implement robust mechanisms to cite sources and allow users to verify information.
  5. 5Gather user feedback iteratively: Continuously refine the system based on the needs and experiences of scientists and clinicians.

Original post by Piotr Grabowski, Mohamed Alameen, Jorge Bretones, Sabina Cardell, Miguel Carmona, Gavin Edwards, Ben Grainger, Sameh Hassan, Erik Jansson, Artur Kuziakhmetov, Albert Maristany, Hebatallah Mohamed, Andriy Nikolov, Sebastian Nilsson, Mark O'Donoghue, James Pacileo, Ashiq Sultan, Alex Voegele, Michael Ughetto

"arXiv:2608.12395v1 Announce Type: new Abstract: We describe Research Assistant, an internal LLM-based system developed at AstraZeneca to help scientists and clinicians explore biomedical questions across a broad range of data sources. The system provides a chat-style interface th…"

View on X

Originally posted by Piotr Grabowski, Mohamed Alameen, Jorge Bretones, Sabina Cardell, Miguel Carmona, Gavin Edwards, Ben Grainger, Sameh Hassan, Erik Jansson, Artur Kuziakhmetov, Albert Maristany, Hebatallah Mohamed, Andriy Nikolov, Sebastian Nilsson, Mark O'Donoghue, James Pacileo, Ashiq Sultan, Alex Voegele, Michael Ughetto on X · view source

Want to go deeper?

Turn these trends into skills with Learnijoy's hands-on AI & tech courses.

Explore courses

More in AI Engineering & DevTools