Creating Intuitive AI Agents for PowerShell..
summary
Shows exactly what is built and how: demos create AI agents inside AI Shell, deploy an Azure OpenAI deployment with Bicep, and wire in providers like Ollama and Copilot in Azure. One demo registers an MCP tool to fetch real PowerShell help so models stop hallucinating examples. A more advanced interpreter agent generates, runs, and self-remediates PowerShell scripts until they stop erroring, demonstrating concrete capabilities, limitations, quota issues, and how to ground models with real data.
AI-written summary · about
chapters (19)
- 0:00Introduction
- 1:15AI agents and components
- 7:36AI agents evolution 2014–now
- 11:08AI agents building blocks
- 16:50Avoiding hallucinations
- 18:16AI agent spectrum
- 20:47Simple agents with Copilot
- 26:44Customizing ChatGPT
- 35:15Model training data
- 42:57Comparing AI models
- 45:18AI Shell introduction
- 50:36Azure OpenAI with Bicep
- 58:06Fine-tuning models
- 1:01:05Configuring AI Shell
- 1:05:22Token usage and pricing
- 1:07:36Advanced AI Shell agents
- 1:10:25Simulating AI team
- 1:20:47Home automation with AI
- 1:29:17Key takeaways and closing thoughts
related talks
39mEvolving Intelligence: Designing Autonomous AI Agents
25mBuilding AI Agents
56mBuilding AI Agents and Multi-Agent Systems with Azure AI
01h 08mFrom Model Integration to Agents: Building Intelligent .NET Applications
48mTurning PowerShell Commands into AI Tools
07h 25mAI Agents Conference 2026 - Day 2
58mAdvanced Autonomous AI Agents: Memory, Collaboration and Evolution in Action
36mBuilding AI Agents in .NET with Azure AI Foundry