
Video unavailable
This video can't be played right now.
Stop wiring infrastructure, start building agents
https://aka.ms/foundry-portal https://aka.ms/InsideMicrosoftFoundryPlaylist Every agent project seems to start the same way containers, scaling, state, auth. Most of the time, you don't need any of it. You rebuild Sparkles as a prompt agent: a model, instructions, and the Cupcake Store MCP tools, defined once and versioned in Microsoft Foundry. No container, no server, no pipeline. Then you hit the edge where a prompt agent isn't enough, such as custom orchestration or your own framework, and move the same agent to a hosted agent. Callers still invoke it by name. By the end, you'll know which one to reach for first, and when to switch. 0:00 - Scale Complex Cupcake Orders 0:34 - Introducing the Copilot Harness 1:07 - Configure Harness Tools 1:38 - Add Reusable Agent Skills 2:44 - Enable Bulk and Party Orders 3:18 - Process Orders from CSV Data 4:25 - Configure the Agent with YAML 4:59 - Define Skills and MCP Tools 5:32 - Scale Without Managing Infrastructure #copilot #harness #microsoft #foundry
Suggested videos

Steve Ballmer at NET Conference going crazy about Developers! | 1999
Economic Archive · 0:57

Microsoft Build 2026 Day 1 LIVE | Opening Keynote, Live Coding & Demos
Microsoft Developer · 8:52:26

From rows to reasoning: Designing databases for AI apps and agents | BRK223
Microsoft Developer · 45:09

SQL and AI for Developers in Action
Microsoft Developer · 59:10

Microsoft Fabric for Power BI developers - 3.5 HOUR FREE COURSE
Learn Microsoft Fabric with Will · 3:29:41

How to Build an App with Azure SQL Using AI Coding Agents | Data Exposed
Microsoft Developer · 10:09

Running a 120B model offline on Surface RTX Spark Dev Box with Intelligent Terminal
Microsoft Developer · 8:28

How does Aspire make Copilot more efficient when I'm building an app?
Microsoft Developer · 1:11