flowchart LR
A["📖 Read Concept<br/>Understand the Why"] --> B["🧠 Study Theory<br/>Build the Mental Model"]
B --> C["🎮 Try Simulation<br/>Experiment Safely"]
C --> D["💻 Complete Lab<br/>Build Something Real"]
D --> E["🚀 Apply It<br/>Solve Your Problem"]
E --> A
Fundamentals of AI
A Practical Guide for the Modern Professional
A hands-on, interactive introduction to Artificial Intelligence for professionals and students. From understanding how large language models work to building production-ready RAG pipelines and AI workflows — no PhD required.
Preface
There is a particular kind of intellectual vertigo that strikes a professor when he realises, mid-lecture, that the subject he is teaching has fundamentally changed — not gradually, not incrementally, but overnight. This book was born in that moment.
The Disruption I Did Not See Coming
For years, I taught Data Analytics to MBA students at Lagos Business School. The course followed a well-worn path: descriptive statistics, exploratory analysis, regression modelling, data visualisation, decision frameworks. It was rigorous, practical, and by most measures effective. Graduates left equipped to interrogate data, build dashboards, and make evidence-driven arguments in the boardroom. I was proud of what we built.
Then came late 2024.
I remember the precise quality of the unease. Students were submitting work that was — and this is the honest word — too good. Not plagiarised, not copied, but transformed. They were producing analyses that would have taken me a week to structure, in an afternoon. They were generating visualisations I would have spent hours polishing, in minutes. They were asking questions in class that were not prompted by the lecture but by conversations they had been having, privately, with AI systems that had already digested the same material I was about to teach.
I sat with that unease for longer than I should have. And then, one evening, I stopped sitting with it and started asking a harder question: Is the course I am teaching still the course my students need?
The answer, when I forced myself to give it honestly, was no.
Not because Data Analytics had become irrelevant — it had not, and never will. But because the workflow of data analytics had been so profoundly disrupted by AI that teaching the old workflow was, in a very real sense, teaching a language that was being translated in real time by a tool sitting in every student’s pocket. The syntax had changed. The vocabulary had changed. The very definition of what it means to be analytically capable had shifted under my feet.
The traditional Data Analytics curriculum assumed a world in which the primary bottleneck was human analytical capacity — the ability to write queries, construct models, interpret outputs, and communicate findings. AI had not eliminated that bottleneck; it had moved it. The new bottleneck was something more fundamental: the ability to ask the right questions, to direct intelligent systems toward the right problems, and to evaluate and deploy the outputs those systems produce. Technical execution — once the hard part — was becoming a commodity. Strategic and conceptual intelligence — once assumed — was becoming the scarce resource.
I had to pivot. Immediately.
A Classroom Became a Laboratory
What began as an emergency curriculum revision became something far more interesting. I did not simply add an “AI module” to the existing Data Analytics course. I started over. I asked my students — sharp, ambitious, sceptical MBA students from across Nigeria and Africa — to help me explore a question neither of us could fully answer: How do you actually use AI to solve real business problems?
The classroom became a laboratory. We experimented with large language models, built retrieval pipelines, connected AI to business data, automated analytical workflows, and designed prompts that could do in seconds what previously required hours of specialist work. We broke things. We discovered limitations. We were surprised — repeatedly — by what was possible.
And as we worked through that exploration, something unexpected happened. A new horizon revealed itself. What had started as a course about data analytics became something considerably larger: a course about how artificial intelligence can leverage data to solve business problems across every dimension of the enterprise.
Not just in the analytics team. Not just in the IT department. But in finance, operations, marketing, human resources, strategy, risk, customer experience — everywhere that decisions are made on the basis of information, which is to say, everywhere. The integrated vision that emerged was not of AI as a tool for data scientists but of AI as infrastructure for organisational intelligence. Every function could be augmented. Every workflow could be rethought. Every professional — not just the technical ones — needed a working understanding of what this technology could and could not do.
That realisation is what this book is.
Why Lagos Business School, and Why Now
Lagos Business School exists to develop leaders who will build Africa’s future. That mission has always been ambitious. It is now, in the age of AI, both more urgent and more achievable than ever before.
Africa stands at a peculiar advantage in this moment. We are not burdened by legacy systems at the scale of older economies. We are not institutionally invested in the workflows that AI is now disrupting. We have a young, digitally fluent population, a tradition of entrepreneurial improvisation, and an extraordinary range of complex problems — in infrastructure, healthcare, agriculture, finance, governance — that represent precisely the kind of challenges that AI-powered solutions can address.
The risk is not that AI will pass Africa by. The risk is that African professionals will become consumers of AI built elsewhere, for problems defined elsewhere, optimised for contexts that are not ours. The professionals who learn to build, direct, and deploy AI systems — who develop the conceptual fluency to work at the level of design rather than merely use — will be the ones who shape how this technology serves African enterprises and communities.
That is the professional this book is written for.
What This Book Is — and What It Is Not
This is not a book for data scientists or software engineers, though they may find it useful. It is a book for the business professional who understands that AI is now part of the operating environment — as fundamental as spreadsheets were in the 1990s or the internet was in the 2000s — and who wants to move from passive familiarity to active capability.
It is a book that takes you seriously. It does not condescend with oversimplifications, nor does it disappear into technical abstraction. It assumes you are intelligent, curious, and pressed for time — and it respects all three.
It is also, unapologetically, a book rooted in practice. Every conceptual chapter is paired with a hands-on lab. Every theory is grounded in a real use case. The interactive simulations embedded throughout are not decoration — they are the course. You learn this material by doing it.
And it is a book built on an honest acknowledgement: I did not know all of this when I began. I learned much of what is in these pages by teaching it — by standing in front of MBA students who demanded that the ideas be clear, practical, and connected to the work they were actually doing. That pedagogical pressure produced a better book than any amount of solitary expertise could have.
A Word on the Pace of Change
I want to be transparent about something that every responsible author in this field must acknowledge: AI is moving faster than any book can track. Some of the specific tools and platforms discussed in these pages will have evolved by the time you read them. New architectures will have emerged. Capabilities that seem remarkable today will seem ordinary tomorrow.
This is not a reason to despair — it is a reason to focus on foundations rather than features. The mental models in this book — how language models process information, how embeddings represent meaning, how retrieval systems connect AI to specific knowledge, how agents coordinate complex tasks — will remain valid long after specific tools have been superseded. A professional who understands the principles can adapt to any new tool. A professional who only knows the tools is perpetually starting over.
Build the foundations. The rest will follow.
How to Use This Book
The chapters are ordered deliberately, each one building on the last. But they are also designed to be self-contained, so that a practitioner who needs to go straight to prompt engineering or RAG implementation can do so without working through every preceding chapter.
Start from the Foundations chapter — From Synapses to Silicon — and read linearly. The history of AI is not an optional detour; it is the conceptual map that makes every subsequent chapter comprehensible.
You may move quickly through the historical foundations and spend the majority of your time on the concept chapters (Parts I–V) and their associated labs. Focus especially on Prompt Engineering and RAG — these are the capabilities with the most immediate business application.
The practice labs are where you will spend most of your time. Each lab is designed to produce a working artefact — a semantic search engine, a RAG pipeline, a stateful AI workflow — that you can adapt to your specific context. The code is provided in both Python and R.
What You Will Be Able to Do
By the time you close this book, you will have moved from understanding AI to building with it. Specifically, you will be able to:
- Explain how large language models work — not just what they produce, but how they produce it
- Make direct API calls to frontier AI models, without intermediary tools
- Build multi-step AI pipelines using LangChain that connect AI to your organisation’s data
- Engineer prompts that extract dramatically better results from any AI system
- Create a functioning semantic search engine — the technology powering enterprise knowledge management
- Implement Retrieval Augmented Generation (RAG), the architecture that allows AI to reason over your specific documents and data
- Design stateful AI workflows using LangGraph that can execute complex, multi-step tasks autonomously
- Understand the Model Context Protocol (MCP) — the emerging standard that will define how AI systems connect to enterprise tools and data sources
Each of these is not an academic exercise. It is a capability you can deploy at work on Monday morning.
A Final Note
The students who sat in my classroom in late 2024 — who pushed me to rethink everything I thought I knew about teaching analytics — deserve the first acknowledgement in this book. Their intellectual honesty was more useful than any dataset.
My colleagues at Lagos Business School, who gave me the institutional latitude to experiment and fail and experiment again, deserve the second.
And you, the reader, who has decided that understanding this technology is worth the investment of your most precious resource — your time — deserve the third.
AI is not magic. It is mathematics, engineering, and data, built by human beings with human limitations and human blind spots. But it is also, genuinely, the most significant amplifier of human capability that our generation has encountered. The professionals who understand it deeply will have an extraordinary advantage over those who merely use it.
This book is your invitation to that deeper understanding.
Let us begin.
Prof. Bongo Adi Lagos Business School April 2026
Prerequisites & Setup
What You Need
- A computer with internet access
- Python 3.10+ or R 4.3+ installed
- An OpenAI API key (free tier works for most labs)
- About 2–4 hours per chapter
Quick Environment Check
# Run this in your terminal to verify your setup
import sys
print(f"Python version: {sys.version}")
# Install core dependencies
# pip install openai langchain chromadb python-dotenv# Run this in R to verify your setup
R.version.string
# Install core dependencies
# install.packages(c("httr2", "jsonlite", "tidyverse", "reticulate"))