Your Name Is On It
The Right Way to Work With AI
AI demos are optimised to impress. Real work is optimised to ship. These are not the same thing, and confusing them is the single biggest mistake I see people make when they start using AI tools.
There are two ways to use AI. One feels fast and exciting at first. The other feels slower and more deliberate. Only one of them actually works when your name is on the output.
The Trap
The speed trap is asking AI to do something large and vague in one shot. “Write me a function that handles user authentication.” It produces something. It looks plausible. But it doesn’t quite fit your codebase, it makes assumptions you didn’t ask for, and when something breaks you’re not entirely sure why.
The speed trap feels productive because output appears quickly. The problem is that speed without understanding is just fast confusion.
But the real cost isn’t bad output. It’s that you can’t stand behind the output. If someone asks you a question about the code, you can’t answer confidently. You can’t ship something you don’t fully own.
The Shift
The leadership approach starts with a different question. Not “can AI do this?” but “how do I work with AI so that I can genuinely stand behind what comes out?”
Think of it like being a tech lead who delegates to a junior developer. You don’t hand over the entire problem and disappear. You break the work into meaningful pieces. You give clear context. You review what comes back. And when you present the final result to a client, you can say with confidence: I know this work, I own this work.
Before I hand anything to AI, I’ve already thought through the big picture, the why, and how the work breaks down. That thinking is mine. AI helps me execute it.
Reviewing means checking, at every level, whether this is even better than what you would have done yourself. Big picture first, then the small stuff.
The leadership approach is slower at first. But it compounds - and before long, you are producing better work faster than the speed trap ever delivered. The more clearly you brief AI, the better the output. The more you review and course-correct, the less you have to fix later.
Leading AI well means the same things leading people well means: give clarity on what good looks like, set high standards, and don’t accept output that doesn’t meet the bar just because it arrived quickly.
The shift is from “AI as a vending machine” to “AI as a capable team member who needs good leadership to do their best work.”
The Same Principle
This plays out beyond code too. Most leaders I work with still build presentations the old way: open PowerPoint, start adding bullets, hope it hangs together by the end. When they try AI, they just do the same thing faster. “Write me a 10-slide deck on our Q3 strategy.” The output looks like a deck. It isn’t their thinking.
The speed trap version: dump the brief into AI, get a deck back, tweak a few words, send it out. It looks fine. But in the room, when someone asks a sharp question, you hesitate. Because you didn’t build it, you assembled it.
The leadership approach to a deck starts before you open any tool. What is the one thing I want this audience to walk away believing? What do I want them to feel? What do I want them to do? Once you are clear on that, AI becomes enormously useful. You are directing it, not hoping it figures out your intent.
A good presentation isn’t a collection of slides. It’s an argument. AI can help you build the argument faster, but only if you already know what you’re arguing for.
The difference shows up in the room. A deck built with the leadership approach, where you drove every decision, is one you can present with confidence, go off-script, handle questions, own the narrative. A deck from the speed trap is one you hope nobody looks at too closely.
The Standard
The standard is simple: would you be comfortable if someone asked you to walk them through every decision in this output? If yes, you used AI well. If you’d have to bluff your way through, you fell into the speed trap.
The goal is not to use AI less. It’s to use it in a way that makes your work genuinely better, not just faster to produce.
Speed is a byproduct of the leadership approach, not the goal. When you brief AI well, review carefully, and hold a high standard, you get great output fast. But you get there because you led well, not because you moved fast.
Every time you put your name on something, you are making a claim: I stand behind this. AI doesn’t change that claim. It just changes how you got there.