The draft is the cheap part

The first draft used to be the hard part of the job. It’s free now, and almost nobody has rethought what to do with the hours that frees up.


The cost moved

An AI tool writes a competent first draft of almost anything in about the time it takes to describe what you want. Copy, a brief, a deck outline, a function, a plan. The making of the thing, the part that used to eat the afternoon, is close to free.

So the job didn’t get easier. It moved. What’s left is the work that was always the harder half, and we just didn’t have to face it this often: deciding whether the thing is any good, and knowing what to do when it isn’t.

Competent is the trap

Here’s the uncomfortable part. AI output is rarely terrible. It’s confidently fine. It reads finished. It has the shape of good work: the right headings, a reasonable structure, sentences that sound like they mean something.

Fine is the dangerous zone. Terrible you catch. Fine slips through, because noticing what’s generic, or quietly wrong, or off for your brand takes a trained eye and a second you’re tempted not to spend. The tool will hand you fine all day. It will never tell you that fine isn’t good enough here.

Judgment is the scarce skill now

What separates people now isn’t who can produce. Everyone can produce. It’s who can look at three competent drafts and know which one is right, which one is a trap, and which one to throw out and start over.

That’s taste. It’s judgment. It’s knowing what good looks like in your context, for your audience, against a standard you carry in your head. None of that ships in the tool. You build it over years of doing the work, getting it wrong, and seeing why. It’s a lot of what an agency year installs and in-house quietly loses.

You can’t license taste, you teach it

Most organizations are about to make the same mistake with judgment that they made with every tool before it. Hand it out, expect quality to rise, and act surprised when the floor comes up but the ceiling doesn’t.

The floor coming up is real, and it’s good. More people can make a passable thing. But the ceiling, the distance between passable and right, is decided by judgment. Judgment is the slowest thing to build and the one nobody puts in the budget. It’s the same story I keep coming back to: the tool is the easy part, the thinking is the hard part, and it’s a teaching problem, not a buying one.


The free first draft is one of the better things to happen to this work in years. It just doesn’t do what people think it does. It doesn’t replace the judgment. It raises how much the judgment matters, because that’s now the only part that’s actually yours. The teams that come out ahead won’t be the ones generating the most. They’ll be the ones who can still tell the difference.



I write about creative leadership, AI adoption, and the parts of the job nobody puts in a keynote. One post every two weeks.