Journal · The AI Architect · 2026-08-03
Break a big goal into small, clear parts before you hand it off. You can't oversee what you haven't divided up.
Break a big goal into small, clear parts before you hand it off. You can’t oversee what you haven’t divided up.
The trouble isn’t that people don’t want to delegate well. It’s that most of us hand off a shape we haven’t looked at closely ourselves. We say “handle the launch” or “get the report done” the way we’d point at a mountain and ask someone to move it. The instruction feels clear in our own head, because we’re holding the whole picture. The person receiving it is holding nothing but the sentence.
This is the real tension: dividing work takes time and mental effort up front, and it’s tempting to skip that step because handing off the whole blob feels faster. It isn’t. It just moves the confusion downstream, where it costs more to fix.
Name the pieces before you name the person
A goal like “improve customer onboarding” isn’t one task, it’s a bundle. Somewhere inside it are decisions about what’s broken, what success looks like, who needs to weigh in, and what order things should happen in. If you hand over the bundle as-is, the other person has to do your thinking and their doing at the same time. That’s two jobs, and most people can only really do one well.
Breaking a goal apart means sitting with it long enough to ask: what are the actual pieces here, and which ones are mine to decide versus theirs to execute? This isn’t about micromanaging every step. It’s about being honest that “improve onboarding” contains at least five or six distinct decisions, and someone has to make each one, even if that someone is you, quietly, before the handoff.
Try this: before you assign anything bigger than a single afternoon’s work, write the goal at the top of a page, then list every sub-decision or sub-task you can think of underneath it, even badly. Don’t organize yet, just dump. You’ll often find the goal wasn’t one thing at all — it was four or five things wearing a single name.
Give each part a shape someone else can hold
Once you’ve broken the goal into pieces, the next mistake is handing over the list as-is, still tangled together. A good piece of delegated work has a few plain features: it’s small enough to picture finishing, it has a clear “done” signal, and it doesn’t quietly depend on three other things nobody’s mentioned.
“Handle communications for the launch” is not a clear part. “Draft the customer email, get it to me by Thursday, keep it under 200 words” is. The difference isn’t length or bossiness — it’s that the second version tells the person exactly what finished looks like. They don’t have to guess what you were picturing.
This is also where oversight becomes possible at all. You can’t check in on “handle communications” in any useful way — there’s nothing concrete to look at. You can check in on a 200-word draft due Thursday. The division isn’t just kinder to the other person; it’s what makes your own oversight actual oversight, instead of vague hovering.
Try this: for each piece you plan to hand off, write one sentence that says what “done” looks like, specific enough that two different people would agree on whether it happened. If you can’t write that sentence, the piece isn’t ready to delegate yet — it’s still your job to define.
Watch what breaking things down teaches you
There’s a quieter benefit to this work that’s easy to miss. When you break a goal into parts, you often discover you didn’t fully understand the goal yourself. Maybe you thought “improve onboarding” mostly meant fixing the welcome email, and it’s only when you list the pieces that you notice the real friction is actually in account setup, three steps later. The division isn’t just prep for handing off — it’s a check on your own thinking.
This means the work of splitting a goal into parts isn’t overhead to get through quickly. It’s where a fair amount of the actual planning happens. Treat it as thinking time, not paperwork, and it’ll pay you back twice: once in a clearer handoff, and once in a clearer sense of what you were actually trying to do.
I’ve spent a long time watching good delegation fail for this exact reason — not from bad instructions, but from whole, undivided ones. The book goes further into how to size these pieces, how to tell which parts genuinely need your judgment versus someone else’s, and what oversight looks like once the dividing is done. But the habit itself starts small: don’t hand off anything you haven’t first broken into parts you could explain to a stranger in a sentence each.
Go deeper. The full method is in The AI Architect. New here? Start with the free companion pack, or explore the series.