Unlock Claude in Microsoft Foundry | ODSP934

Caroline Matthews and Keiji Kanazawa introduce how to use Anthropic Claude models in Azure AI Foundry, including a live demo, guidance on choosing Claude models, and what it looks like to deploy and monitor agentic workflows inside existing Azure infrastructure.

Overview

This Build 2026 session covers Claude models being available in Microsoft (Azure) AI Foundry and how enterprise teams can move from pilot to production using Foundry features for model selection, evaluation, and observability.

What the session focuses on

• Deploying frontier models (Claude) within existing Azure infrastructure via Azure AI Foundry
• Building agentic applications (agents and an agentic loop)
• Using Foundry capabilities to support production readiness:
  • Evaluation
  • Observability
  • Monitoring and playback for agent workflows
• Getting started considerations, including billing via MACC

Speakers

• Caroline Matthews
• Keiji Kanazawa

Chapters (from the video description)

• 0:00 - Introduction to Microsoft and Anthropic Webinar on Building Smarter Agents
• 00:10:26 - Streamlined path from pilot to production with unified evaluation and observability
• 00:11:13 - Introduction of Claude model lineup in Azure AI Foundry
• 00:14:10 - Cupcake Factory demo – deploying and coding with Claude Sonnet in Foundry and VS Code
• 00:19:52 - Introduction to Agent Concept and Agentic Loop
• 00:22:00 - Demo of Running Claude Agent in Microsoft Foundry
• 00:31:39 - Playback and Monitoring Features for Agent Workflow in Foundry
• 00:32:11 - Understanding the Role of Evaluations in Agentic Applications
• 00:36:07 - Conclusion and Resources for Building Agents with Claude and Foundry

Event context

This is a Microsoft Build 2026 on-demand session. More sessions are available at https://build.microsoft.com