Book 4 of 10 · Capable Beginner tier
The one skill that makes every AI tool obey.
+ add the Intermediate Vault pack for £27.00 at checkout
The difference between a mediocre AI answer and a brilliant one is almost never the tool — it's the prompt. This is the deep, practical guide to prompting: showing the model what good looks like, grounding it in your own source material, controlling tone and format, and building a reusable library so you never start from a blank box again.
It goes well beyond the basics covered in Book 1: giving AI examples of the output you actually want, feeding it your own documents so it stops guessing, and controlling exactly how it reasons through a problem step by step. There's a full chapter on building yourself a personal library of prompts you reuse rather than reinventing each time.
It closes with the failure modes worth knowing — the specific ways prompts go wrong, and how to prompt safely when what you're asking for actually matters.
Anyone who has used AI and thought "this could be so much better" — and wants it to be, without becoming a programmer.
The kind of week each book is built to produce.
E-commerce customer support
a growing support inbox with slow response times. Moved to AI-drafted replies that a human reviews before sending.
This pattern generally cuts response times substantially while keeping a person in control of what actually gets sent.
Recruitment or a hiring-heavy business
screening a large volume of applications by hand. Moved to a structured AI screening pass against a scored rubric, with a human making the final call.
This tends to make shortlisting faster and more consistent, without removing human judgement from the decision.
No — Book 1 covers the basics needed to get started day one. This is the deeper craft: grounding, format control, reasoning, and a reusable library, for someone already comfortable with AI day to day.
Yes — Chapter 10 is specifically about prompting for automations, where a prompt has to work reliably without a human checking every output.
No. It's still plain-English and tool-agnostic — the skill is in how you phrase instructions, not in any code or API work.
