The book
Don’t believe in AI. Specify it.
A practitioner’s guide for Product Owners and Business Analysts. One running case, Atlas, goes from a vague request for “an AI agent” to a governed pilot, and you learn every decision along the way.
- Ebook
- PDF + EPUB
- Price
- £14.95
- Paperback
- 7 × 10 in · £19.95
ISBN 978-1-0654535-5-0 (paperback) · 978-1-0654535-0-5 (ebook)
The ebook downloads instantly as a PDF and an EPUB. The Kindle edition is on Amazon. Paperback £19.95 on Amazon. Every enterprise training seat includes a copy.
Is it for you?
Written for the people in the middle.
You've been told to “use AI”
and need to turn that into a problem, a user and a measure of success.
You write the requirements
and need to specify behaviour, sources, boundaries and failure, not just features.
You'll face a governance board
and need a proposal with evidence, controls and a clear decision route.
What’s inside
Thirteen chapters. One lifecycle.
The chapters follow a real enterprise lifecycle, from shared vocabulary to a capstone Canvas, so each one builds on the last.
Plus appendices with the Enterprise AI Solution Canvas, a canonical artefact map, the five decision routes, a glossary and references to the EU AI Act, NIST AI RMF, ISO/IEC 42001 and OWASP GenAI guidance.
- The vocabulary of AI, GenAI and LLMs
- Where AI creates real value
- Building the business case
- Shaping a bounded use case
- Discovery that tests the idea
- Writing requirements for AI behaviour
- Data and knowledge readiness
- RAG, assistants and agents
- Risk, controls and governance
- Delivering a governed pilot
- Evaluation you can defend
- Running AI through its lifecycle
- The Enterprise AI Solution Canvas
From the author
Why I wrote it.
I have watched a three-month build collapse at sign-off because “agent” meant four different things to four teams. The problem was never the model. It was that nobody had written down what the AI should do, for whom, using which sources, and where it had to stop.
This book is the method I wish those teams had. It is practical, it uses one case all the way through, and it is honest about when the right answer is to park or stop.
Madhusudhan Konda
Read the book? Write a review →
Try before you buy
The free Canvas and the Atlas worked example come straight from the book.
Who is this book for?
Product Owners, Business Analysts, product managers and delivery leads who have to turn an AI ambition into a product that can be built, governed and tested. It is written with regulated and document-heavy organisations in mind, such as banks, insurers and the public sector.
Do I need a technical background?
No. You do not need to code or know how models are trained. The book explains the terms you need, such as LLMs, RAG, assistants and agents, in business language.
What will I learn?
How to separate real AI opportunities from demo-shaped ones, build a business case on evidence, write requirements for systems that do not always give the same answer, judge whether your data is ready, choose between search, RAG, assistants and agents, govern the risk, evaluate beyond a single accuracy score and run a pilot that ends in a clear decision.
Is it tied to a particular AI vendor or tool?
No. The method works whatever platform your organisation uses. It references current standards and guidance, including the EU AI Act, the NIST AI RMF, ISO/IEC 42001 and OWASP GenAI.
Where can I buy it?
The ebook (PDF and EPUB) is £14.95 from this site. The Kindle edition is on Amazon. Paperback £19.95 on Amazon. Every enterprise training seat includes a copy.
What are the ISBNs?
Paperback: 978-1-0654535-5-0. Ebook: 978-1-0654535-0-5. Published by ChocolateMinds, first edition 2026.