AGENTIC WORKFLOWS 6 HOUR COURSE: Beginner to Pro (2026)

簡介：

In this 6-hour Agentic Workflows course (Beginner to Pro, 2026), you will learn how to design, build, and deploy real business automations using modern LLM tooling and frameworks like DOE, Claude Skills, and MCP, with hands‑on demos from lead scraping and proposal generation to CRM managers and Gmail workflows. We cover the full loop from planning and tool use to memory, reflection, self‑annealing reliability, and scaling via parallelization, then finish with cloud and scheduled deployments so you can run workflows autonomously day‑to‑day. Chapters 00:00:00 Introduction 00:01:37 Foundational concepts 00:04:35 Scraping leads with Agentic Workflows 00:11:51 How Agentic Workflows have changed the game (automation tools & chatbots vs Agentic Workflows) 00:30:03 The 5 steps (planning, tools, memory, reflection & orchestration) 00:49:59 The evolution of interfaces (text vs GUIs) 00:52:51 The issue of variability (probabilistic systems vs deterministic work) 01:07:09 Using LLMs vs Python scripts 01:09:58 Integrated Development Environments (IDEs) 01:27:45 Antigravity walkthrough & generating proposals with Agentic Workflows 01:32:45 VS Code walkthrough & running UpWork scraper 01:34:54 Project & agent workspaces: structure & setup 01:43:00 Sales call transcript to generated proposal 01:45:41 Directive Orchestration Execution (DOE Framework) 02:10:03 Building first agentic workflow using DOE 02:18:26 Building a CRM manager for ClickUp 02:28:02 Claude Skills 02:32:04 Building with Claude Skills 02:35:56 Model Context Protocol (MCP) & input context windows 02:53:25 MCP in action (Gmail) 03:01:01 Systematic approach to building workflows (prompts, files, self-annealing, autonomy) 03:08:19 Building a LinkedIn lead generation scraper 03:23:17 Building an improved lead generation scraper using parallelization 03:33:26 How to improve Agentic Workflows over time 03:41:01 Self-annealing workflows 03:50:30 AI safety & text interface 03:57:43 Maximizing efficiency & productivity (speaking, specificity, questions, pasting over typing) 04:03:43 How to effectively utilize API documentation 04:09:37 Why you should watch workflows as they run & handling longer workflows 04:11:56 Setting up hooks (ex: sound notifying to check output when workflow finishes) 04:15:16 Reviewing output & chaining workflows together 04:28:45 Putting workflows in the cloud & webhook deployment 04:50:51 Scheduled deployment 04:54:35 Scaling with parallelization of multiple agents 05:00:34 Using agentic workflows in day-to-day 05:14:18 Sub-agents 05:40:31 Outro