googleadsagent.ai
English · Français · Español · 中文 · Nederlands · Русский · 한국어
From Absolute Zero to Deploying Multi-Agent Systems
A Step-by-Step Tutorial for Every Experience Level
A free, open-source, 42-page crash course on building AI agents — written by John Williams, Senior Paid Media Specialist at Seer Interactive with 15+ years managing $48M+ in digital advertising spend, and creator of googleadsagent.ai.
Every concept explained with accessible analogies. Every step verified with "What You Should See" callouts. Every day ends with a playable video game built through natural-language prompting. Formatted in APA 7th Edition with 16 formal references.
- Anthropic Academy Integration — 13 free courses with certificates linked throughout, including Claude 101, Claude Code in Action, MCP courses, AI Fluency, and Agent Skills
- Interactive MCQ Quizzes — 6 graded quizzes (21 total questions) replace text-area assessments with instant feedback and scoring
- XP & Gamification System — Earn XP for completing days and quizzes, level up from Novice to Master, earn 6 achievement badges
- Daily Streak Tracking — Return daily to build your streak; 3-day streaks unlock badges
- New Products Coverage — Claude Code, Claude Cowork (Jan 2026), Agent Skills, MCP updates
| Section | Contents |
|---|---|
| Cover | Branded title page with author credentials |
| Table of Contents | 5-page manual TOC with all 87+ headings |
| Preface | Why This Course Exists, Audience & Paths, How to Use, Acknowledgments |
| Executive Summary | 1-page overview for quick reference |
| Section 0 | Complete A-Z setup: Claude, ChatGPT, Cursor, Kaggle, GitHub, Python |
| Days 1–5 | Each with: Learning Objectives → Content → Key Takeaways → Self-Assessment |
| Appendix A | Complete Resource Library (white papers, courses, frameworks, platforms) |
| Appendix B | Glossary of 22 key terms |
| Appendix C | googleadsagent.ai Technical Reference (28 actions, 6 sub-agents) |
| References | APA 7th Edition bibliography (16 entries) |
| About the Author | Full bio and credentials |
| Path | Who | Focus | Time/Day |
|---|---|---|---|
| 🟢 Observer | Executives, strategists | Concepts only | 2–3 hrs |
| 🔵 Builder | Marketers, analysts | Hands-on labs + games | 4–6 hrs |
| 🟣 Architect | Developers, engineers | Deep theory + production | 6–8 hrs |
| Day | Topic | White Paper | Game |
|---|---|---|---|
| 0 | Complete Environment Setup | — | — |
| 1 | Introduction to AI Agents | Introduction to Agents | 🎮 Tetris |
| 2 | Agent Tools & MCP | Agent Tools & MCP | 🎮 Zelda Adventure |
| 3 | Context Engineering | Sessions & Memory | 🎮 Football (Tecmo Bowl) |
| 4 | Agent Quality & Observability | Agent Quality | 🎮 Super Mario Bros |
| 5 | Production & Multi-Agent | Prototype to Production | 🎮 Duck Hunt |
| Cap | Everything combined | All five | 🎮 AI Game Master RPG |
googleadsagent.ai is woven throughout as a living reference implementation:
| Component | Implementation | Day |
|---|---|---|
| Brain | Claude by Anthropic (Opus 4.5 + Sonnet 4.5) | 1 |
| Tools | 28 Python API actions with live Google Ads API access | 2 |
| MCP | Adapter layer: tool_executor.py |
2 |
| Context | CEP Protocol + Session State Manager | 3 |
| Quality | Top-Down Reporting + CONFIRM write-safety | 4 |
| Multi-Agent | 🦁 Simba · 🐠 Nemo · ❄️ Elsa · 🧞 Aladdin · 🌊 Moana · 🤖 Baymax | 5 |
| Production | CLI, REST API, Docker, Python SDK | 5 |
| Format | Description | Link |
|---|---|---|
| 📄 Word (.docx) | Full course with formatted callouts, tables, code blocks | docs/AI-Agents-Crash-Course-V4.docx |
| Print-ready, 42 pages | docs/AI-Agents-Crash-Course-V4.pdf |
|
| 📝 Markdown | Read on GitHub | docs/COURSE.md |
Each day includes a game built entirely through natural-language prompting. Copy the prompt, paste into Claude or Cursor, play instantly.
| Day | Game | Prompt |
|---|---|---|
| 1 | Tetris | games/day1-tetris/prompt.md |
| 2 | Zelda Adventure | games/day2-zelda/prompt.md |
| 3 | Football (Tecmo Bowl) | games/day3-football/prompt.md |
| 4 | Super Mario Bros | games/day4-mario/prompt.md |
| 5 | Duck Hunt | games/day5-duckhunt/prompt.md |
| Cap | AI Game Master RPG | games/capstone-rpg/prompt.md |
All sources formally cited in the course:
| Source | Date |
|---|---|
| Sundar Pichai — Google I/O 2025 Keynote | May 2025 |
| Demis Hassabis — Gemini 2.0 Launch | Mar 2025 |
| Google — Agent2Agent Protocol (A2A) | Apr 2025 |
| Google — Cloud Next 2025 | Apr 2025 |
| Sundar Pichai — NRF 2026 Remarks (UCP) | Jan 2026 |
| Alphabet Inc. — Q2 2025 Earnings | Jul 2025 |
| Google/Kaggle — 5 White Papers | 2024–2025 |
| Anthropic — MCP Documentation | Nov 2024 |
| Yao et al. — ReAct Paper | 2022 |
| Karpathy — Vibe Coding | Feb 2025 |
| Microsoft — AI Agents for Beginners | 2025 |
Full bibliography in resources/REFERENCES.md
All 13 courses are free with certificates on Anthropic Academy (Skilljar):
| Course | Focus | Level |
|---|---|---|
| Claude Code in Action | AI coding assistant, GitHub, MCP servers, hooks, SDK | Intermediate |
| Claude 101 | Chat, Cowork, Code modes, Projects, Artifacts, Skills | Beginner |
| AI Fluency: Framework & Foundations | The Four Ds — Delegation, Description, Discernment, Diligence | All levels |
| Building with the Claude API | API calls, streaming, function calling, structured outputs | Developer |
| Intro to Model Context Protocol | MCP servers/clients in Python, tools, resources, prompts | Developer |
| MCP: Advanced Topics | Sampling, notifications, file system, transport mechanisms | Advanced |
| Introduction to Agent Skills | Build and share SKILL.md workflows in Claude Code |
Developer |
| AI Fluency for Educators | AI Fluency in teaching and institutional strategy | Educators |
| AI Fluency for Students | AI skills for learning and career planning | Students |
| AI Fluency for Nonprofits | AI fluency for mission-driven organizations | Nonprofits |
| Teaching AI Fluency | Teaching and assessing AI Fluency in instructor-led settings | Instructors |
| Claude with Amazon Bedrock | Anthropic models on AWS | Cloud |
| Claude with Google Cloud Vertex AI | Anthropic models on Google Cloud | Cloud |
The live course includes:
| Feature | Details |
|---|---|
| MCQ Quizzes | 6 quizzes, 21 questions, instant grading with feedback |
| XP System | Earn XP for completing days (20 XP each) and quizzes (30-50 XP each) |
| 6 Levels | 🔰 Novice → 📘 Learner → 🔧 Builder → ⚡ Practitioner → 🏗️ Architect → 👑 Master |
| 6 Badges | 🎯 First Quiz · 💯 Perfect Score · 🏆 All Days · 🔥 3-Day Streak · 🎓 Academy · ⭐ Completionist |
| Daily Streaks | Visit daily to build your streak counter |
| Progress Tracking | All progress saved locally in your browser |
White Papers — All 5 Google/Kaggle Papers · Intro to Agents
Anthropic Academy — Claude Code in Action · Claude 101 · All 13 Courses
Protocols — MCP Docs · A2A Docs
Courses — Microsoft AI Agents · Hugging Face · DeepLearning.AI
Tools — Claude · ChatGPT · Cursor · googleadsagent.ai
ai-agents-crash-course/
├── README.md ← You are here
├── LICENSE ← MIT
├── docs/
│ ├── AI-Agents-Crash-Course-V4.docx ← Full course (Word)
│ ├── AI-Agents-Crash-Course-V4.pdf ← Full course (PDF)
│ └── COURSE.md ← Full course (Markdown)
├── games/
│ ├── day1-tetris/prompt.md
│ ├── day2-zelda/prompt.md
│ ├── day3-football/prompt.md
│ ├── day4-mario/prompt.md
│ ├── day5-duckhunt/prompt.md
│ └── capstone-rpg/prompt.md
└── resources/
├── REFERENCES.md ← APA bibliography
└── LINKS.md ← All links & resources
John Williams — Senior Paid Media Specialist at Seer Interactive with 15+ years managing $48M+ in digital advertising spend across Google Ads, Meta, Microsoft, and Amazon. Former NortonLifeLock (192% YoY growth), Gen Digital, Avast, Farmers Insurance. Creator of googleadsagent.ai. Speaker at Hero Conf. Assistant football coach at Casteel High School. Former Washington State University football player (2002–2005).
GitHub · LinkedIn · X · Substack · itallstartedwithaidea.com
MIT License. Share it. Teach it. Fork it. The future of AI agents belongs to everyone.
Version 5.0 · March 2026 · Free & Open Source
googleadsagent.ai/course · googleadsagent.ai · It All Started With A Idea