How Should Executives Learn AI? A Practical Upskilling Path

So here's what I keep seeing with the executives I coach. They want to get good at AI, so they sign up for a course, or they watch a stack of demos, or they read every think piece that crosses their feed. Months go by. They know a lot about AI now. They still can't actually use it on a Tuesday afternoon when there's a real decision in front of them. They studied the instrument without ever picking it up.

I want to be clear up front about what this article is and isn't. I've written separately about what most corporate AI training programs get wrong, so I'm not going to relitigate that here. This is the other side of that coin. This is the practical path for one leader, you, who actually wants to get good at this. What to learn, in what order, and how to practice so it sticks.

Why Most Leaders Learn AI the Wrong Way

The most common way executives try to learn AI is by consuming content about it. A webinar here, a demo there, a long article on the train. It feels productive because you're absorbing information, and at the end you can talk about AI more confidently in a meeting. The trouble is that watching someone else use a tool well teaches you almost nothing about using it yourself, in the same way that watching a guitarist on YouTube doesn't put a single callus on your fingers.

Getting good at AI is a doing skill, not a knowing skill. The leaders I've watched get genuinely fluent didn't read their way there. They picked something real, used the tool on it, got a mediocre result the first time, figured out why, and tried again. That loop, the trying and adjusting and trying again, is the entire game. You can't shortcut it by watching someone who's already past it.

Here's the test I'd apply. If you've spent ten hours "learning AI" this month and zero of those hours were you doing your own real work with the tool open, you haven't been learning AI. You've been learning about AI. Those are different activities, and only one of them makes you better.

Start With One Real Task From Your Own Week, Not a Course

So if courses aren't the move, where do you start? With a task you already have to do. Look at your actual week and find one thing that eats your time and recurs: the board update you rewrite every month, the first draft of a strategy memo, prepping for one-on-ones, working through a hiring decision, turning messy notes into something coherent. Pick one. That's your training ground.

The reason this beats a course is that the stakes are real and the feedback is honest. When you use AI on a made-up exercise, you don't really care if the output is good. When you use it on the board update you're actually sending, you care a lot, and that pressure is what teaches you. You'll notice immediately when the output is generic, when it missed the point, when it sounds nothing like you. That noticing is the lesson.

Don't try to learn ten tools at once either. Pick one general-purpose assistant and stay with it. I use Claude every day and do most of my input by talking instead of typing, but the specific tool matters far less than the commitment to use the same one on real work until it's second nature. One tool, one recurring task, every day. That's the whole starting kit.

The Thing That Actually Matters Is Context, Not Prompt Tricks

There's a whole cottage industry of prompt tricks out there, magic phrases that supposedly make AI ten times better. Most of it is noise. The real thing that separates useful AI from useless AI is context, meaning how much the tool actually knows about your business, your goals, your standards, and the situation you're in.

Think about onboarding a sharp new hire. On day one they're brilliant in general and useless to you specifically, because they don't know your company, your customers, your history, or how you like things done. You don't fix that with a clever instruction. You fix it by sitting them down and explaining the business. AI is the same. When you give it a one-line prompt with no background, it answers the way it would for anyone on earth, which is to say generically. When you tell it who you are, what you're trying to do, what good looks like, and what to avoid, it starts producing work that's actually yours.

So as you practice on that first real task, the skill to build isn't memorizing prompt formulas. It's learning to brief the tool the way you'd brief a capable assistant. What's the goal, who's the audience, what's the context, what does done look like. Get good at that and you've got most of what matters. This is also the skill that transfers across every tool, every model, every update, because it's really just clear thinking and clear communication.

A Practical Order to Learn In

People want a curriculum, so here's the order I'd actually follow. Three stages, and you don't move to the next one until the current one is a habit.

Stage 1: Use it on real work.

This is the first few weeks, and it's the most important. One tool, one recurring task, daily. Brief it well, judge the output honestly, adjust, repeat. The goal here is simply to get reps and to build the reflex of reaching for AI on real tasks instead of defaulting to the way you've always done it. Don't optimize anything yet. Just play, on real work, every day.

Stage 2: Build reusable context, a second brain.

Once you've got a few wins, you'll notice you keep re-explaining the same things: your company, your role, your audience, your preferences. That's wasted effort. The next skill is saving that context once so the tool already knows it. Most tools let you store standing instructions or reference documents the assistant can draw on every time. Think of it as building a second brain the AI can read from. You write your business down once instead of re-typing it into every conversation. This is the step that turns AI from a clever toy into something that genuinely saves you hours.

Stage 3: Automate what you repeat.

Only after the first two are solid does automation make sense. By now you'll see the tasks you do the same way over and over, and many tools let you package a repeated workflow so it runs on command instead of from scratch each time. This is where leaders who've put in the reps start getting real time back on their calendar. But it's the last stage on purpose. Automating a process you don't yet understand just lets you make the same mistake faster.

How to Build the Habit

None of this works if it's a once-a-month effort, which brings me back to the instrument. Nobody gets good at guitar by practicing for four hours on the first Saturday of the month. They get good with twenty minutes most days. AI is identical. A little, daily, on real work, beats a big push every now and then, because the skill is really a set of reflexes and reflexes only form through repetition.

The simplest way I know to build the habit is to make a rule that for one specific task, you always start with AI in the loop. Not every task, just the one you picked. When that one becomes automatic, add a second. You're not trying to transform your whole workflow overnight. You're trying to wear a groove, one task at a time, until reaching for the tool feels as natural as opening your email.

And give yourself permission to be bad at it for a few weeks. The executives who quit usually quit in week one, when the output is mediocre and it feels slower than just doing it themselves. That phase is real and it's temporary. It's the same awkward stretch as the first weeks on any instrument, where it sounds worse before it sounds better. Push through it and the curve bends fast.

Where a Coach or Workshop Speeds It Up

You can absolutely learn this on your own. The path above is the whole thing, and plenty of leaders walk it without help. What a coach or a hands-on workshop does is cut the trial and error. Left alone, most people spend the first month picking the wrong first task, briefing the tool poorly, and getting discouraged before they hit their first real win. A good guide gets you to that win faster, points you at the task most likely to pay off for your specific role, and shows you the context-building step before you've wasted weeks re-explaining yourself.

The way I think about it is the same as any other skill where a coach helps. You could learn an instrument from videos, and some people do, but a teacher watching what you actually play will fix in one session what would've taken you two months to notice on your own. That's most of what working with me one-on-one is: I sit with you on your real work, we pick the right first tasks, and I help you build the context and habits so the skill sticks instead of fading after the initial excitement. For a team that needs to get fluent together, a workshop does the same thing at scale.

However you do it, with help or on your own, the path doesn't change. Real work, reusable context, then automation. Reps over courses. A daily habit instead of a heroic push. Get those right and you'll be good at AI in a couple of months, while the leaders still collecting webinars are exactly where they were when they started.