Morning friend ☕️

This is the last week of The Builder's Workshop series.

For four Mondays, we've been putting the room together. The door. The bench. The shelves. The crew.

Today, I want to show you what the room is actually for.

Welcome to Part 5: The Builder.

If this is your first week here, the workshop is simply a way of organising AI around one real project. Mine happens to be the first Scary Management book and my workshop has really helped me be more effective and efficient.

How?

Grab your coffee ☕️

Let's chat!

This was before the workshop existed.

I had the book in front of me, but before I could safely add to it, I needed to give the system everything around it to double check I wasn’t “breaking” anything.

The doctrine. The canon. What had already been decided. Almost like regression testing.

I put it in by hand each session, checking as I went. By the time I finished, I had used up most of my time and hadn't added anything to the manuscript.

This is the bit I had underestimated.

A serious project creates its own supporting work. The “grunt” work. You have to find things, remember why you made an earlier choice, compare versions, prepare a useful brief, check the result and leave a record for next time.

As the book grew, so did the work around it. My backpack grew steadily heavier.

AI could help me lighten that load. And let me focus my precious energy and attention where it mattered.

Sitting down to write that earlier evening bothered me for another reason.

I had meant to work on the book. I spent the session maintaining everything around it.

Back in May I wrote about Lebo, the friend in the novel who refuses to play along.

Thandi sends her a long voice note late one night, hoping to be told she's doing alright. Lebo sends back:

"I didn't hear you in any of it. Not your voice. Not your laugh. Where did you go?"

She was talking about how Thandi’s job was affecting her.

But I could recognise the question in how my book project was going.

On the 8th of July, two chapters had been rewritten and checked by the workshop.

The answer from the workshop was: they were good.

I read the work and asked a slightly annoying question:

It's good. Is it great?

That question went onto the bench beside the chapters.

The shelves already held the book's voice rules, earlier decisions, continuity and the things those chapters needed to pay off later. I didn't have to assemble that material again before the question could be useful.

Then the crew went back through the work.

The story pass found six possible upgrades across the two chapters. One of them caught a character completing a thought too early. A similar idea returned six chapters later, where it was meant to land properly. Leaving the earlier version alone would have weakened the later one.

The continuity check traced the change across both chapters. The reader check asked whether the scenes merely worked or whether somebody would actually feel them. The existing rules stopped the answer becoming "add more words and hope".

All of that came back to me with the reasons attached.

I read it, approved the six changes and owned what happened to the book.

The door made the attempt affordable. The bench held the chapters and their history. The shelves supplied the context. The crew prepared, challenged and checked the work.

My job was to decide that good was not yet the standard, and then decide what deserved to change.

The supporting work now has places to live between sessions and a crew to execute on it.

I could have tried to hold the story, the continuity, the reader experience, the voice rules and the decision history in my head at once. On that earlier evening, that is what I had been doing.

Now the workshop can prepare much of it and put it in front of me.

Some of the crew's contribution is creative. It can suggest language, restructure a section, find alternatives and make an argument I had not considered. Organising context can change what I understand about the work too.

But the room does not know what I want this book to become.

It cannot set the standard unless I give it one. It cannot decide which trade-off I am willing to make, or carry the responsibility when I make the wrong call.

I still do that part.

Maybe you want to start something new.

Maybe there is a process inside your existing business that works only because you are there to remember all the exceptions.

Maybe you already have a half-built project, scattered across old chats, documents and the notes app on your phone.

You do not need fourteen chapter files or seven roles to begin.

Start with a four-line commission:

I am building:
It is for:
What I know that should shape it:
The decision I need help seeing clearly:

For a two-person design studio, it might look like this:

I am building: a proper quoting process.
It is for: me, and the freelancer who quotes when I'm away.
What I know that should shape it: vague scopes and friendly clients are where we lose money.
The decision I need help seeing clearly: whether we should keep offering fixed-price work next year.

Put that page in one permanent folder. Add the notes and rules the project needs. Give the same commission to a builder and a sceptic. Put their answers next to each other, make the call yourself and write down why.

That is enough workshop for a first project.

And when you come back, the work is waiting for you with a little more of its memory intact.

What is one project you would like to spend more time building and less time administering?

Give it ten minutes. Put its name at the top of a page and complete the four lines.

Keep the scope to one piece of real work. Commission it and give it somewhere to wait for you.

Five weeks ago, this series began in a corridor, in front of a door many of us had stopped trying.

The room on the other side turns out to be fairly ordinary. A folder. A page of what you know. Two roles that argue. And you, deciding.

I've still got a book to finish.

If this brought one of your own projects to mind, hit reply and tell me what you're commissioning. I read every one.

The workshop can't want it for you.

But if you do, the door is open.

Until next Monday, friend ☕️

Vaugan

This series gives you the mental model. I’m also building a step-by-step Builder’s Workshop Guide to help you assemble a working version around one real project.

And you can reuse this workshop structure for any project you create with your AI tool.

It begins with one folder and the four-line commission, then walks through the bench, shelves, first roles and the decision loop. The practical pieces are included: file structure, prompts, role templates, decision records and the checks that stop the room quietly drifting away from what you meant.

The first rung needs no code and no elaborate tool stack. The aim is that you finish with a working workshop that has your project in it, rather than a folder full of mine.

I'm also including the mistakes I made along the way because those are part of the wiring too.