Part 4: Prompt Engineering and AI Tools
Working With AI, Not Just About AI
"The interface between human intent and machine capability is where all the value lives."
Parts 1 through 3 taught you to understand AI — its foundations, its algorithms, and its most powerful techniques. Part 4 shifts the focus from understanding AI to using AI. The distinction matters enormously.
A business leader who understands machine learning but cannot prompt an LLM effectively, evaluate an AutoML platform, or architect an AI-powered workflow is like an executive who understands financial theory but has never opened a spreadsheet. Knowledge without application is an expensive luxury.
Part 4 is where theory becomes practice.
You will learn to craft prompts that reliably extract value from large language models — not through trial and error, but through systematic engineering. You will build multi-step AI workflows that combine retrieval, reasoning, and generation. You will evaluate no-code AI platforms, navigate cloud AI services, and deploy AI in marketing and customer experience — the domain where most organizations encounter AI first.
What You Will Learn
Chapter 19: Prompt Engineering Fundamentals establishes the discipline of prompt engineering — prompt anatomy, zero-shot and few-shot techniques, role-based prompting, iterative refinement, and the PromptBuilder class that makes prompt construction systematic rather than ad hoc.
Chapter 20: Advanced Prompt Engineering introduces sophisticated techniques: chain-of-thought reasoning, tree-of-thought exploration, prompt chaining for complex tasks, structured outputs, and the PromptChain class that orchestrates multi-step LLM interactions.
Chapter 21: AI-Powered Workflows moves beyond single prompts to complete systems. Retrieval-Augmented Generation (RAG), vector databases, AI agents, and workflow orchestration come together in a working RAG pipeline that connects Athena's knowledge base to its customer service operation.
Chapter 22: No-Code / Low-Code AI examines the platforms that democratize AI — AutoML tools, drag-and-drop model builders, and embedded AI features. You will learn when these tools are genuinely useful, when they create shadow AI risks, and how to evaluate vendors with clear criteria.
Chapter 23: Cloud AI Services and APIs navigates the hyperscaler AI landscape — AWS, Azure, and Google Cloud AI services. Pricing models, vendor lock-in risks, and multi-cloud strategies will help you make infrastructure decisions that affect years of organizational capability.
Chapter 24: AI for Marketing and Customer Experience applies everything from Parts 1–4 to the domain NK Adeyemi knows best. Personalization, chatbots, customer journey analytics, attribution modeling, and the ethical boundary between helpful and invasive — the "creepy line" — are explored through NK's capstone project at Athena.
The Athena Story Continues
Part 4 bridges Athena's Scaling Phase and Deepening Phase. The organization's AI ambitions are growing, and new tools — LLMs, RAG systems, no-code platforms — promise to accelerate progress. But each new tool introduces new decisions. Should Athena build its own RAG pipeline or buy a platform? Should business analysts use AutoML tools, and if so, what governance is needed? When does a chatbot help customers and when does it frustrate them?
NK's internship at Athena becomes central to the narrative. Her marketing AI project — building a personalization system that respects customer boundaries — earns her a return offer and sets up her trajectory toward the Director of AI Strategy role she will eventually hold.
The Tools Evolve Fast. The Principles Don't.
The specific AI tools and platforms discussed in Part 4 will evolve. Some vendors named here will be acquired, some products will be discontinued, and new ones will emerge. This is inevitable in a rapidly moving field.
What will not change: the principles of systematic prompt engineering, the architecture patterns for AI workflows, the evaluation criteria for vendor selection, and the organizational dynamics of AI tool adoption. Learn the principles. Apply them to whatever tools exist when you need them.
Let's get to work.