April 7, 2026
The Prompt Isn't the Play Call. It's the Practice Rep.
What coaching youth baseball taught me about AI — and why the most important skill right now has nothing to do with technology.
By Jim Edgett

Most of the kids who struggle early aren't struggling with the skill. They're struggling with confidence — and what nobody tells you is that confidence doesn't come from encouragement alone. It comes from knowing what right feels like.
I've had a few players over the years who couldn't field a ground ball cleanly. Different kids, same problem: glove stopping just short — three, maybe five inches off the ground — and the ball skipping right underneath. You'd almost miss it watching from the dugout. The kid thinks he's doing it right. His glove is near the ground. But near isn't down, and that gap is all a baseball needs.
I'd tell them: get the glove all the way down. They'd try. Same result. I'd show them with my hands. They'd nod. Same result.
It wasn't stubbornness. It just feels genuinely strange to press your glove into the dirt — uncomfortable, almost wrong. Without ever experiencing what "all the way down" actually feels like, they had nothing to reach toward.
So I stopped coaching the error and started coaching the image.
I walked out to shortstop and had my assistant coach feed me ground balls — easy ones first, then harder. I didn't narrate. I just played: feet squaring up, body dropping, eyes on the ball, glove pressing into the dirt before the ball even arrived. I let them watch until something clicked. Not the mechanics. The picture. A real one. A felt one.
Then I called them out.
Glove down on the first rep. Every time.
Coaching, at its best, is building a player from the inside out. You're not fixing an error. You're building a vision — something clear enough that the player can find their way back to it on their own.
I've been thinking about those kids a lot lately. Because I see the same thing playing out with AI.
Something Remarkable Is Happening. And It's Okay to Feel Two Things About It.
We are living through a genuine shift. Skills that used to take years to develop — writing, research, analysis, coding, design, financial modeling — are now accessible to almost anyone with a laptop and a clear enough question.
That is not hype. It is real. And it is worth pausing to appreciate.
A first-generation college student can now produce research that rivals a consultant's. A small business owner can write marketing copy that reads like it came from an agency. A team of five can move with the reach of a team of fifty.
The barrier to access has collapsed.
And that raises a fear that's just as real: what happens to me in a world where the tool can do what I spent years learning?
It's a fair question. I won't dismiss it. But I think sitting with the fear leads to the wrong conclusion.
The people who feel most displaced by AI won't be the ones who know the most. They'll be the ones who never develop their judgment — who accept the first output because they don't have a strong enough picture of what better looks like. Who hand the tool a vague question and take whatever comes back.
The people who thrive will be the ones who bring what no model has: lived experience, organizational context, taste, and a clear vision of the outcome they're building toward. That's not a soft skill. That's leverage.
The coach who gets extraordinary results from their players doesn't get replaced by a better playbook. They become more valuable because they know how to use one.
Learning to get great results from AI isn't protecting yourself from a threat. It's building the skill that matters most right now. And the good news — the genuinely encouraging news — is that it's learnable. Here's where to start.
Show the Tool What You're Reaching For
The biggest leap I made wasn't a better command or a smarter workflow. It was learning to demonstrate before I asked.
Most prompts are instructions: write me a professional summary, give me three options, make this sound better. The AI responds — and what comes back is technically fine and somehow completely off. Generic tone. Wrong shape. It sounds like an AI, not like you.
That's not a tool problem. That's a vision problem. You know what you want, but you haven't shown the tool what it looks like.
The fix is simple: paste an example first. A paragraph you wrote. A piece of writing that sounds the way you want to sound. A previous output you actually liked. Then say: write in this style.
The quality shift is immediate. Not because you found a magic phrase, but because you gave the tool a picture to work toward. You became the coach at shortstop.
The First Output Is the First Rep
Coaches don't draw up a play and expect perfection the first time. The play is the framework. The reps are where the learning happens.
Same with AI.
The first response isn't the answer. It's the opening attempt. You read it the way you'd watch a player's first swing — what's off, what's close, what do I want more of? Then you coach it forward.
"Keep the opening — rewrite everything after the second paragraph." "Too formal. Write it like I'm talking to a colleague." "Good structure. Now cut it in half."
That follow-up prompt is the real work. The first response just shows you where to start.
And this is the thing I most want you to hear: that is not failure. That is the method. No coach apologizes for running a drill twice. Iteration is how good becomes great. The willingness to stay in the loop — to keep coaching instead of accepting the first rep — is what separates excellent output from adequate output.
Build It Layer by Layer
Great coaches don't teach the full system in week one. Footwork first. Then the transfer. Then the throw. Then game pressure. Each layer builds on something solid underneath.
The same patience pays off when you're working toward something complex — a document, a strategy, a presentation. First prompt: establish the structure and voice. Second: build out the core. Third: refine and push further.
Each good output becomes the foundation for the next ask. You're not just generating content. You're building something that reflects your thinking at every step — because you stayed in the loop, coaching toward your vision the whole way.
The Vision Was Always Yours
Back to those kids at shortstop.
They got good. Not because I had magic words, and not because the drill was clever. We built a loop — show what right looks like, take a rep, correct it, go again. Eventually the picture on the inside matches what's happening on the outside. That's when confidence stops being fragile. That's when the skill belongs to them.
Here's what I want to leave you with.
The tools have arrived. The access is real. The fear is understandable. But the answer to the fear isn't to stay on the sideline — it's to step into the loop.
No AI has your experience. Your instincts. Your understanding of your customers, your team, your organization. No model knows what you're trying to build or why it matters. That part is yours. It always will be.
The work of this moment isn't learning more commands. It's developing your eye — getting clear enough on what right looks like that you can close the gap when you're three inches short. Building the confidence to keep coaching until the output is actually good, not just close.
That's not a technical skill.
It's the most human skill there is.
And it's exactly what a kid learned one afternoon with his glove in the dirt.
Journey Gain works with operators and leaders who are building AI-enabled revenue systems — and who want to do it with real organizational judgment behind it. If this landed, reach out at jim@journeygain.com.
Jim Edgett
Jim Edgett is the founder of Journey Gain, which builds AI-enabled identity and loyalty systems for QSR and retail operators. He has spent 20+ years at the intersection of loyalty, first-party data, retail media, and CX — including GameStop’s 65M-member loyalty ecosystem, Salesforce/IBM engagements with Dick’s Sporting Goods and TaylorMade, and advisory work with multi-location restaurant and retail brands.