AI 專精訓練營：從零到代理架構師
一個為期五天的密集訓練營，旨在將初學者轉變為人工智能代理架構師。課程內容涵蓋提示工程的「BRIC」框架、閱讀與病毒式文案撰寫的內容加速技巧、辦公室自動化（Excel 與 PowerPoint）、使用 RAG（檢索增強生成）建立「第二大腦」，以及打造自主數位員工。
課程總覽
📚 內容摘要
一個為期五天的密集訓練營，旨在將初學者轉化為人工智慧代理架構師。課程涵蓋提示工程的「BRIC」框架、閱讀與病毒式文案創作的內容加速技巧、Excel 與 PowerPoint 的辦公室自動化、利用 RAG（檢索增強生成）技術打造「第二腦」，以及建立可自主運作的數位員工。
從基礎提示工程到部署能自動化複雜辦公流程的自主代理，一份全面掌握人工智慧的指南。
🎯 學習目標
- 提示工程精通：應用 BRIC 框架、工具選擇策略，以及少樣本提示等進階技巧，消除通用化的人工智慧輸出。
- 內容加速：掌握「內容煉金術」，瞬間總結冗長文檔，並使用風格基因提取技術生成具有人味、具病毒傳播力的文案。
- 辦公室自動化：透過人工智慧在數分鐘內生成複雜的 Excel 公式與完整的 PowerPoint 簡報，徹底擺脫重複性勞動。
- 知識管理：運用 RAG 技術建立「第二腦」，解決幻覺問題，並讓人工智慧安全地查詢私人資料。
- 代理架構：設計、建構並部署可執行多步工作流程、與即時網路互動的自主「數位員工」。
課程 共 5 课时 · 预计 20.0h
課程
Lesson
This lesson teaches you to optimize AI performance by matching specific tasks to the right model architecture and using the BRIC framework to provide essential structural context. You will also learn how to use few-shot prompting to refine the AI's output style, ensuring your results are both accurate and tailored to your professional needs.
This lesson introduces the Content Alchemy Framework, a methodology for rapidly transforming dense information into strategic intelligence by prioritizing high-speed synthesis and human-centric stylization. Students learn to move beyond passive reading by using the Info Juicer process and recursive prompting to extract actionable insights while maintaining a unique, professional voice.
This lesson explores the shift from manual, syntax-heavy office tasks to intent-based productivity, where AI acts as a translator for complex Excel formulas and automated slide deck creation. By adopting an architectural mindset, you will learn to leverage natural language to direct AI agents, allowing you to focus on high-level business logic and strategic storytelling rather than technical execution.
This lesson introduces Retrieval-Augmented Generation (RAG) as a method to eliminate AI hallucinations by grounding responses in your own curated documents rather than relying on internal training data. You will learn how to transition from prompt engineering to source engineering using tools like NotebookLM to build a reliable, verifiable "Second Brain" for personal and enterprise knowledge management.
This lesson explores the transition from passive chatbots to autonomous agents by integrating planning, memory, and tool-based execution into a recursive workflow. Students will learn how to design multi-step processes that enable AI to interact with real-world data and self-correct through sequential logic.