I was standing on the 4th tee box when the first offer came in.
My phone buzzed with a Facebook notification. Someone wanted the golf cart. Then another message. Then a third. Three offers within hours of posting—for a cart I'd built a complete sales website for while playing nine holes with my Dad, my brother, and two of my sons.
We were out there with both carts that day. The new one we'd just given Dad for Christmas. And the old Chiefs-themed Club Car we were trying to sell.
Here's what I'm still processing: I didn't miss a single shot because of the website project. Didn't step away to "handle something real quick." The AI workflow happened in the gaps—between holes, while others were putting, during the natural pauses that exist in any round of golf.
This is what I keep discovering about AI: it's not about doing more. It's about fitting things into the spaces that already exist.
The Christmas Gift That Created a Problem
Dad's an avid golfer. Plays almost daily. Lives on the course—literally, his house backs up to the fairway. His old Club Car DS had served him well, but it was showing its age.
So for Christmas, we upgraded him. New cart. The kind of gift that matters when golf is your daily routine.
But now we had a problem: what to do with the old one.
My boys and I spent an afternoon with Dad fixing it up. We washed and detailed it, touched up the paint, and cleaned the undercarriage. And we installed brand-new batteries—six Pro-Guide 6V deep-cycles that cost about as much as some people's car payments.
Those batteries changed everything about how to sell this cart.Why Batteries Are the Whole Game
If you've ever shopped for a used electric golf cart, you know the first question: "How old are the batteries?"
It's the #1 buyer objection. Battery replacement runs $900-1,200. Most used cart listings either dodge the question or bury it. Buyers are rightfully suspicious.
We'd just eliminated that objection entirely. Brand-new batteries, December 2025. The most expensive maintenance item—handled. A buyer could drive this cart for years before thinking about batteries again.
That's a story worth telling. A Marketplace post wouldn't do it justice.
Building the Site While Building Memories
I didn't plan to build a website that day. I'd been noodling on the idea, had some research done, but figured I'd tackle it later.
Then we were on the course, the conversation was easy, and I had my phone. So I started.
The workflow:
ChatGPT handled the research. I fed it the serial number (A90382295S), photos of the cart, and photos of the batteries. It identified the model (Club Car DS, early 1990s), confirmed the aluminum frame (no rust—an important selling point), and helped me think through pricing strategy.
ChatGPT wrote the copy. Headlines, bullet points, the FAQ section, and the whole landing page structure. I gave it the details; it gave me marketing language that emphasized benefits over features.
Replit Agent built the actual site. This is where it got interesting. I basically had a conversation with the AI agent, iterating through maybe 15 rounds of changes in about 90 minutes total.
The Replit conversation looked like this:
"Create me a landing page to promote this golf cart sale."
The first version appeared in about two minutes. Not bad, but the colors were wrong.
"Use a gray tone but with Kansas City Chiefs highlights to accent it."
Better, but it interpreted Chiefs colors wrong—included dark blue, which isn't even a Chiefs color.
"Remove the dark blue. Use yellow, red, and white from the Chiefs brand."
Now we're talking. Chiefs Red (#E31837), Gold (#FFB81C), clean white background.
Then I uploaded the actual photos. Changed the phone number. Uploaded the real battery image. Asked it to optimize for mobile. Asked it to optimize for speed. Asked what else would make it better.
That last prompt surprised me. Replit suggested scroll animations, a lightbox gallery, and a value comparison card showing "$0 Future Cost" versus "+$1,200 Risk" for carts without new batteries. Features I hadn't thought of. I said yes to three of them.
Between the 3rd and 4th holes, the site was live.
Where AI Got It Wrong
I'm not going to pretend this was flawless. AI made mistakes, and I almost published them.
The voltage error: ChatGPT confidently identified the cart as a 48V system based on serial number research. Sounded right. I didn't question it.
The actual batteries? Six 6-volt batteries. 6 x 6 = 36V. Not 48V.
This is AI's blind spot with physical verification. It can decode serial numbers and make educated guesses, but it can't look at the actual batteries sitting in front of you. I caught the error during final review, but it could have been embarrassing with a knowledgeable buyer.
The temporal language problem: The initial copy said "batteries purchased yesterday, installed today." That's fine for the day you write it. It's nonsense a week later. I had to manually flag that for correction—changed it to "purchased brand new in December 2025."
The Facebook cache: After correcting the voltage on the site, the Facebook link preview still showed "48V" because social platforms cache Open Graph data. I boosted the post before refreshing the cache. Rookie mistake. The preview still shows the wrong voltage as of this writing.
The phone number: Somehow, during Replit iterations, the wrong phone number got entered. The original ChatGPT research had the correct number. Somewhere in the back-and-forth, a different number appeared. I caught it, but not before a version went live briefly.
These aren't polished lessons. They're mistakes I made two days ago. I'm sharing them because someone reading this will make the same ones if I don't.
What Actually Happened
Posted Christmas Eve, around 3 PM. Boosted the post for local reach.
Within hours: three offers. One serious enough to call it "official."
As of right now, the cart hasn't been sold. My rule: it's not sold until it's paid for. But the interest validated the approach. A professional landing page with a clear value proposition stood out from the typical Marketplace listings.
Whether we close at $3,000 or negotiate down, the battery investment is protected. The story we told—"brand new batteries, turn-key, no surprises"—resonated.
The Part I Keep Thinking About
I built a complete sales website while playing golf with three generations of my family.Not instead of being present. Not by stepping away to "handle work real quick." During the natural gaps that exist in any group activity.
My boys saw me on my phone occasionally. They also saw me sink a putt on 7 and hook one into the trees on 8. They'll remember the round with Grandpa and Uncle more than they'll remember Dad checking his phone.
That's the thing about AI that I keep discovering. It's not about productivity. It's not about efficiency. It's about fitting things into spaces that already exist without expanding those spaces into the time that matters.
Dad got a new cart for Christmas. The old cart is generating offers. And I didn't miss a single moment with my family to make either of those things happen.
If You Want to Try This
The workflow is reproducible:
Use ChatGPT for research and copy. Feed it whatever details you have—serial numbers, photos, specs. Let it help with pricing strategy and marketing language.
Use Replit Agent for the build. It's conversational. You don't need to know code. Just describe what you want and iterate.
Verify physical details manually. AI can't see your actual product. Double-check specs against reality.
Refresh social caches before boosting. Use Facebook's Sharing Debugger if you change anything on the site after initial posting.
Keep the scope small. Single-item sale? You don't need a backend, database, or contact form. Static site + phone number is optimal.
The whole project took about two hours. It could be done in less if you're not on a golf course between iterations.
The Landing Page Is Still Live
You can see the actual site here: gardner-golf-cart.replit.app
Fair warning: it might be taken down once the cart sells. But as of this writing, it's live, Chiefs-themed, and waiting for the right buyer.
If you know anyone in the Gardner, Kansas area looking for a turn-key golf cart with brand-new batteries, send them that link. Dad would appreciate it.
And if you try building something similar—a landing page for a sale, AI-assisted while doing something that matters more—I'd genuinely like to hear how it goes.
This is part of my ongoing documentation of AI experiments across business and life. I test these tools so you don't have to guess what works. Sometimes they work beautifully. Sometimes I publish the wrong voltage. Both are worth sharing.
Created with ❤️ by humans + AI assistance 🤖