Choosing and Deploying Advanced AI Models Using Microsoft Foundry
Microsoft Events presents expert guidance on using Microsoft Foundry to evaluate and deploy advanced AI models from providers like OpenAI, Anthropic, and Cohere, featuring speakers William Flannery, Keiji Kanazawa, Naomi Moneypenny, Steve Sweetman, and Tao Zhang.
Choosing and Deploying Advanced AI Models Using Microsoft Foundry
Overview
This advanced session from Microsoft Ignite 2025 explores how Microsoft Foundry brings together the world’s leading AI models—such as Anthropic Claude, OpenAI, Cohere, Meta, xAI, and more—into a unified, secure platform on Azure. The session demonstrates practical approaches for evaluating, routing, and deploying models to optimize for cost, latency, and accuracy, while maintaining enterprise-grade security and responsible AI governance.
Speakers:
- William Flannery
- Keiji Kanazawa
- Naomi Moneypenny
- Steve Sweetman
- Tao Zhang
Key Topics Covered
- Introduction to Microsoft Foundry and unified AI framework
- Secure, scalable deployment of multiple models
- Chapter highlight: Tools for global collaboration, including RF3 for protein modeling
- Latest model integrations
- Anthropic Claude models now available in Foundry
- Cohere models integrated via direct Azure connection
- Access to providers like Black Forest Labs, Mistral AI, Meta, xAI
- Model Router and API Tools
- How the Foundry router intelligently directs requests to the most suitable model based on cost, latency, and accuracy requirements
- API utilization and automation examples
- Building General-Purpose AI Agents
- Leveraging multimodal models for versatile AI agent applications
- Workflow architecture and deployment patterns
- Governance and Responsible AI
- Ensuring enterprise-grade security for every workload
- Built-in governance practices for responsible usage
- Scaling AI Adoption
- Approaches for organizations to scale AI model use with variety and robust deployment options
Additional Resources
Session Chapters
- 0:00: RF3 Tool and Global Collaboration
- 2:33: Microsoft Foundry Introduction
- 5:25: Anthropic Claude Models Announcement
- 15:13: Model Router and API Tools
- 19:08: Building General-Purpose Agents
- 23:56: Model Overview with Naomi Moneypenny
- 24:42: Cohere Integration Announcement
- 30:49: Scaling AI Adoption
- 34:26: Secure, Scalable Deployment with Foundry
Summary
This session provides practical guidance for Microsoft developers and AI practitioners on making intelligent model choices and orchestrating multi-model solutions on Foundry. It emphasizes actionable techniques for integration, deployment, and ongoing management, illustrated with real-world scenarios and expert insights.