Unleash the power of AI: Automate tasks with OpenAI and Semantic Kernel
summary
Shows how Semantic Kernel and .NET hook real LLMs to code: it wires GPT-4o via Azure OpenAI assistants into an event-driven process layer on Dapper/Dapr, uses threads to share conversation state across agents, and adds plugins (deterministic function calls) so an editor agent reliably emits approve/reject events. Demonstrates a writer+editor joke workflow that loops and converges, claims this scales to hundreds or thousands of agents, and notes costs and logging/debugging tradeoffs while the demo runs successfully.
AI-written summary · about
chapters (18)
- 0:00AI automation with OpenAI
- 1:04Automate manual business processes
- 2:06Multi-agent collaboration with AutoGen
- 4:14Break monolithic agents into teams
- 5:18What is a process in Semantic Kernel
- 8:33First process without AI
- 13:35Add AI to workflow step by step
- 18:33Workflows beat standalone chatbots
- 21:03Multi-agent workflows with reviewers
- 27:01Shared state with Azure OpenAI Assistants
- 31:15Agents in OpenAI dashboard
- 34:00Deterministic AI with plugins
- 37:15Prevent infinite loops in workflows
- 41:42Debug agent interactions with thread history
- 45:18Agents maintain state across workflows
- 46:05Azure OpenAI Agents (upcoming)
- 50:14Optimize costs with smaller AI models
- 52:32Summary and resources
related talks
01h 02mDevelop AI agents with Semantic Kernel
01h 00mBuilding your own AI Agent using Semantic Kernel
26mSemantic Kernel: Teaching AI to Not Just Repeat, But Complete!
59mOrchestrate your AI with Semantic Kernel
26mAI powered automation & multi-agent orchestration
46mBuilding AI solutions with Semantic Kernel
46mInfusing your .NET Apps with AI: Practical Tools and Techniques
47mAI Orchestration with Semantic Kernel