Microsoft Developer: Speed up Research with Microsoft Discovery

Microsoft Discovery coordinates specialized AI agents that analyze research evidence, propose candidate materials, run scientific tools, and validate the results, while the scientist sets the goals and makes the key calls. Because it works as a continuous loop, the plan shifts toward the most promising candidates as new evidence arrives.

Ali Zaman, Product Manager at Microsoft, uses battery design as the example: a battery that charges faster, lasts longer, and uses safer materials. He explains why simple search and summarization fall short when evidence comes from research papers, proprietary data, simulations, and lab experiments, and how Discovery mirrors the scientific method to reach results with fewer experiments.



Resources

- Microsoft Discovery https://azure.microsoft.com/en-us/solutions/discovery

- Blog: Microsoft Discovery: Advancing agentic R&D at scale https://azure.microsoft.com/en-us/blog/microsoft-discovery-advancing-agentic-rd-at-scale/

- What is Microsoft Discovery? https://learn.microsoft.com/en-us/azure/microsoft-discovery/overview-what-is-microsoft-discovery

- Explore more Azure resources! https://azure.com/AzureEssentials



Chapters

0:00 Why building a better battery takes more than search

0:56 Why the battery research space grows so fast

1:14 How Microsoft Discovery coordinates AI agents for research

1:35 Why scientists stay in control of AI-driven research

1:50 How Microsoft Discovery works as a continuous research loop