The AI Initiative Playbook

Enterprises are adopting AI quickly. Very few can point to a business outcome that actually changed.

This book is about why.

It is written for CIOs, transformation leaders, enterprise architects, and coaches — the people making AI decisions inside real organisations, not the people building the tools. It will not teach you to build anything. It will not hand you an AI strategy. It will not go deep enough into the technology to replace an engineer.

What it gives you instead is a way of thinking about AI decisions that outlasts whichever model or vendor you are using next year.

Some of what is in here:

• AI amplifies the system you already have. Good systems get better. Broken systems get more visibly broken, faster.
• Capacity freed does not become customer value on its own. It becomes a saving, unless someone deliberately redirects it.
• Let AI narrate. Let code compute. Keep the two apart and both stay checkable.
• Governance intensity should match consequence. Treating it as all-or-nothing is why most agent programmes fail.
• Every task AI takes over was also somebody's practice. The reviewer who stops drafting stops noticing what they used to catch.

Every claim in this book came from building something and watching what happened. Six projects sit behind it, including a governed AI delivery pipeline for regulated delivery and an agentic system with explicit governance boundaries. Where the evidence stops and the reasoning begins, the book says so.

You will also find things that did not work. A governance gate built wrong, and caught by asking a question the tests never asked. An experiment where AI made the writing faster and worse at the same time. Most AI writing reports only the wins, which is why so much of it sounds confident and helps so little.