A field guide to working with AI

Parenting Your AI

A practitioner's guide to raising AI skills with values, not just instructions.

This series follows the full arc of building responsible AI skills: why values matter more than rules, what your AI does not know that you need to teach it, how to build skills that do not break, and how to diagnose them when they fail. Twenty-one essays, written from inside a live enterprise AI system.

14 essaysRead in order or follow your curiosity

The curriculum

Start your journey

The Problem

2 essays

Before you build more capable AI, understand the quiet failures that happen when power arrives before judgment.

01Your AI has all the power and none of the wisdom. That's your problem to fix.02Rules only cover what you anticipated. Values cover the rest.
Begin this chapter

The Character

2 essays

The limits, humility, and self-awareness that give an AI skill a recognisable shape.

01The most important thing you can teach your AI is what it will never do02You're responsible for decisions your AI made while you weren't looking
Begin this chapter

The Craft

4 essays

A thoughtful chapter in the journey toward better AI practice.

01How to write a skill that doesn't break02How to make your AI think, not just answer03The six values I encode into every AI skill I build04How to keep an AI skill working six months after you built it
Begin this chapter

The Field

2 essays

What these ideas look like in real teams, real systems, and the places where trust has to be earned.

01What I wish someone had told me before I started building AI skills at work02What senior AI engineers actually do differently, and what the rest of us can learn from it
Begin this chapter

The Practice

4 essays

Small, repeatable habits for directing, correcting, and organising AI so better work becomes easier to make.

01The fastest way to make AI better at your work is to tell it when it’s wrong02Stop adding instructions. Fix the prompt.03Give examples. It is the most underrated thing you can do with AI.04Your prompt is not the intelligence. It's the signal.
Begin this chapter