Embracing frontier R&D with Microsoft Discovery | DEM315

John Link and Viktor Veis introduce Microsoft Discovery and the idea of “Agentic Discovery,” where autonomous AI agent teams help run end-to-end scientific R&D loops. They demo how the system turns a research question into a generated notebook, analyzes container vs serverless cost and utilization patterns, and produces an executive-ready summary of findings.

Overview

Microsoft Discovery is presented as a lightweight entry point for “Agentic Discovery”: an operating model where autonomous teams of AI agents, guided by human expertise, execute full cycles of the scientific method.

Key ideas covered

What “Agentic Discovery” means in practice

• An operating model for AI-driven R&D loops, not just adding AI tools to existing research workflows.
• Autonomous AI agent teams support:
  • Hypothesis formation
  • Experimentation
  • Reasoning over results
  • Producing outputs for both technical and non-technical audiences

Microsoft Discovery app as an entry point

• The session describes the launch of the Microsoft Discovery app as a lightweight way to start using the system.

Demo flow: from research question to analysis artifacts

• The presenters show a detailed research-question instruction provided to the Discovery system.
• The system generates a notebook and includes analysis elements such as:
  • Cost analysis
  • Duty cycle insights

Architecture and cost analysis example

The demo includes analysis comparing:

• Containers vs serverless utilization
• Cost variation by time of day

Based on repeated analysis, the system recommends a hybrid execution model.

Communicating results to non-technical stakeholders

• The demo includes generating a one-page executive summary intended for a non-technical audience.

Session context

• Microsoft Build 2026 session: DEM315
• Level: Intermediate
• Theme: Cloud platform & data
• Speakers: John Link, Viktor Veis
• More sessions: https://build.microsoft.com