Artificial intelligence is not gradually permeating finance — it is restructuring it at a pace that has no historical precedent. Trading decisions that once required a human analyst and a morning now require a machine learning model and a microsecond. Credit decisions that once required a loan officer and three weeks now require a gradient-boosted tree and an alternative data pipeline. Risk management that once required a weekly committee meeting now requires a neural network and a real-time dashboard.
These are not distant possibilities. They are happening now, and the displacement they produce is real. As AI absorbs the routine cognitive work that once sustained millions of finance careers, the workers it displaces will turn to entrepreneurship as their most viable path forward. Past experience suggests that half will fail. Understanding why, and what AI itself does to that failure rate, is no longer an academic question. It is one of the defining social challenges of the coming decade.
This book explains how these systems actually work, what they actually optimise, and what happens when they are deployed in contexts their designers did not anticipate.
Structured as six self-contained essays, the book addresses core questions across AI in finance: Agentic AI and high-frequency trading; supervised machine learning and credit scoring; unsupervised learning and customer segmentation; reinforcement learning and robo-advisory; and AI-driven financial inclusion.
Essays 1 to 5 follow a consistent three-part structure: the business problem AI addresses, the technical toolkit it deploys, and the ethical and governance dimensions that purely technical accounts invariably obscure. The result is a treatment that is rigorous without being inaccessible, and critical without being dismissive.
Essay 6 departs from this structure deliberately. Written in research article format, it develops an original framework — the Liability Convexity Amplification Effect — to explain why poor microentrepreneurs fail at rates that cannot be attributed to irrationality or poor financial literacy, and why AI credit scoring, even when technically excellent, is likely to deepen rather than dissolve the trap.
Each essay is followed by future research questions suitable for Bachelor's, Master's, and PhD theses, and by review questions that test not only your understanding of the chapter but also your capacity to think beyond it.
Written for finance students, practitioners, and educators who want to understand AI at the level of mechanisms rather than headlines, this book is also the product of a genuine experiment in Human-Agent Collaboration: a sustained dialogue between the author's decades of research in microfinance, financial inclusion, and fintech, and the analytical capabilities of contemporary AI systems.
If you are entering finance today, this is the industry that actually exists. This book is your map.
