Uploaded June 2026 | Updated September 2026, 2 weeks ago
As organizations move from AI experimentation to production deployment, they need platforms that can support the full lifecycle of agent development—from creation and hosting to evaluation, optimization, and governance.
In this session, Sebastian from Microsoft demonstrates how Microsoft Foundry helps teams build, deploy, manage, and continuously improve AI agents at enterprise scale. Through a live demo, he showcases how developers can generate agents with GitHub Copilot, deploy them as hosted services, distribute them through Microsoft Teams and Microsoft 365, and use automated evaluation and optimization workflows to improve performance over time.
The talk also highlights Microsoft’s latest announcements around hosted agents, agent optimization, agent frameworks, tool integrations, voice interfaces, observability, and evaluation tooling. Learn how Microsoft is building a complete platform for enterprise agent development and what it takes to operate production-grade AI systems at scale.
Timestamps
00:00 Introduction to Microsoft Foundry
00:37 The New Era of AI Agents
01:00 Building Production-Grade Agents on Foundry
01:53 Core Components of the Foundry Platform
02:34 Live Demo: Building an Agent in VS Code
03:02 Generating Agents with GitHub Copilot
03:57 Creating a Vendor Management Agent
04:30 Deploying Hosted Agents to Foundry
05:22 Configuring Models and Infrastructure
05:55 Testing Agents in the Foundry Playground
06:18 Publishing Agents to Microsoft Teams
07:03 Closing the Loop with Agent Optimization
07:22 Introducing the Agent Optimizer
08:03 Running Automated Agent Experiments
08:28 Evaluations, Rubrics, and Performance Scoring
09:09 New Microsoft Foundry Announcements
09:26 Agent Harnesses and Agent Frameworks
09:41 Toolboxes and Tool Integrations
09:55 Voice Interfaces for AI Agents
10:15 Hosted Agents and Agent Service Updates
10:25 Scheduled Agent Routines
10:52 Tracing, Evals, and Procedural Memory
11:14 Enterprise Capabilities and Governance
11:48 Getting Started with Foundry
Key Takeaways
• Building AI agents is becoming dramatically easier through AI-assisted development workflows and agent-generation tools.
• Microsoft Foundry provides a complete platform for agent development, deployment, hosting, governance, and observability.
• Hosted agents allow teams to focus on agent behavior while Microsoft manages the underlying infrastructure and runtime.
• Continuous optimization using automated evaluations, rubrics, and experimentation is becoming a core part of the agent lifecycle.
• Enterprise AI platforms increasingly need integrated support for deployment, security, observability, evaluation, distribution, and governance.
--
#MicrosoftFoundry #AIAgents #GitHubCopilot #AgentEngineering #EnterpriseAI #LLMOps #AIObservability #AgenticAI #MicrosoftAI #ArizeObserve
🔗 Try Arize AX & Phoenix OSS: arize.com
🔔 Subscribe for weekly content on LLMs, agents, and evaluation: youtube.com/@arizeai?sub_confirmation=1
As organizations move from AI experimentation to production deployment, they need platforms that can support the full lifecycle of agent development—from creation and hosting to evaluation, optimization, and governance.
In this session, Sebastian from Microsoft demonstrates how Microsoft Foundry helps teams build, deploy, manage, and continuously improve AI agents at enterprise scale. Through a live demo, he showcases how developers can generate agents with GitHub Copilot, deploy them as hosted services, distribute them through Microsoft Teams and Microsoft 365, and use automated evaluation and optimization workflows to improve performance over time.
The talk also highlights Microsoft’s latest announcements around hosted agents, agent optimization, agent frameworks, tool integrations, voice interfaces, observability, and evaluation tooling. Learn how Microsoft is building a complete platform for enterprise agent development and what it takes to operate production-grade AI systems at scale.
Timestamps
00:00 Introduction to Microsoft Foundry
00:37 The New Era of AI Agents
01:00 Building Production-Grade Agents on Foundry
01:53 Core Components of the Foundry Platform
02:34 Live Demo: Building an Agent in VS Code
03:02 Generating Agents with GitHub Copilot
03:57 Creating a Vendor Management Agent
04:30 Deploying Hosted Agents to Foundry
05:22 Configuring Models and Infrastructure
05:55 Testing Agents in the Foundry Playground
06:18 Publishing Agents to Microsoft Teams
07:03 Closing the Loop with Agent Optimization
07:22 Introducing the Agent Optimizer
08:03 Running Automated Agent Experiments
08:28 Evaluations, Rubrics, and Performance Scoring
09:09 New Microsoft Foundry Announcements
09:26 Agent Harnesses and Agent Frameworks
09:41 Toolboxes and Tool Integrations
09:55 Voice Interfaces for AI Agents
10:15 Hosted Agents and Agent Service Updates
10:25 Scheduled Agent Routines
10:52 Tracing, Evals, and Procedural Memory
11:14 Enterprise Capabilities and Governance
11:48 Getting Started with Foundry
Key Takeaways
• Building AI agents is becoming dramatically easier through AI-assisted development workflows and agent-generation tools.
• Microsoft Foundry provides a complete platform for agent development, deployment, hosting, governance, and observability.
• Hosted agents allow teams to focus on agent behavior while Microsoft manages the underlying infrastructure and runtime.
• Continuous optimization using automated evaluations, rubrics, and experimentation is becoming a core part of the agent lifecycle.
• Enterprise AI platforms increasingly need integrated support for deployment, security, observability, evaluation, distribution, and governance.
--
#MicrosoftFoundry #AIAgents #GitHubCopilot #AgentEngineering #EnterpriseAI #LLMOps #AIObservability #AgenticAI #MicrosoftAI #ArizeObserve
🔗 Try Arize AX & Phoenix OSS: arize.com
🔔 Subscribe for weekly content on LLMs, agents, and evaluation: youtube.com/@arizeai?sub_confirmation=1
