The old saying goes, "There are many ways to skin a cat." If there is one thing I’ve learned from working with AI, it’s that this holds completely true in the tech world, too. If you’ve tried to get AI to do anything for you recently, you’ve probably ended up on YouTube, digging through Reddit, reading SubStack’s, or just playing around to see if you can make it work. Ultimately, you find a workaround, and you get it done.
While tutorials and forums have occasionally given me the exact answer I needed, the real breakthroughs usually happen when I'm just messing around and trying different approaches. When you’re building with AI, you have to be prepared to get your hands dirty.
By that, I mean you might have a beautiful vision in your head of what your product or package will do, right up until the moment you start sharing it with your "AI digital team." Whether you're creating a mock-up, a wireframe, a slide deck, or even writing code for scripts and dashboards, what you’ve sketched out in your mind doesn’t always translate directly to your new AI developers.
The Good, the Bad, and the Loops
My journey has definitely had its fair share of frustrating moments. I’ve been caught in loops where the AI keeps spitting back the exact same bad results over and over, completely missing the point. Even worse is when you’ve done four weeks of solid work, and the AI starts the next session by asking a question about something you settled back in week one. Suddenly, you spend the entire day just bringing it back up to speed before you can even move forward.
I’ve even had situations where I enjoyed a couple of weeks of absolute brilliance—fantastic coding, excellent UI—and then, out of nowhere, the AI slowly begins to completely screw it up. There is nothing more frustrating than an AI telling you it can't remember, cannot access its memory, or has no record of past conversations, suggesting you just "start over." Ultimately, you learn to work through this and find workarounds that suit your style. Persistence is definitely key.
I’ve been caught in loops where the AI keeps spitting back the exact same bad results over and over, completely missing the point
But the other side of that coin is pure magic. There are sessions that run perfectly. The ideas flow, the code comes out exactly as you envisioned, and when you run a smoke test or an eyeball check, it works flawlessly. I’ve had many days like that, and it is an incredibly gratifying feeling
Why I Built My Own CMS
"Vibe coding" might be a massive buzzword right now, but in my case, it’s the absolute truth. My ideas flow onto the page, become code, and then transform into a working application. I've gone from building a simple dashboard and an online calculator to a working prototype web app that helps people plan meals and design menus. Now, I've built a full, working CMS (Content Management System) that does exactly what I need for my own websites.
"Why build your own CMS when WordPress works perfectly fine?" I got a lot of pushback on that. You only have to do a quick Google search to find a sea of naysayers asking that exact question. But in my case, I spent years trying to force plugins to work within the rigid WordPress criteria, only for them to fall short of what I actually needed. By building my own base CMS first and adding my own modular platform on the backend, the site runs significantly faster for visitors, and I can build the exact plugins I want to use. So far, I’ve been really happy with the results. For now, vibe coding is exactly where I’m staying.
Mapping the Customer Journey
Right now, I already have a custom analytics platform that tracks outbound links, cross-linking between content, read depth, and the exact journey a customer takes through the site. I built cookie policies that allow users to opt out completely to respect their privacy. But once they consent, the tracking begins, and we can see the story of how they engage with the site, which feeds directly into a reporting dashboard I can view.
Once a visitor opts in via a form—whether for a newsletter or to give feedback—they become a user. That’s when more powerful analytics take over. We can physically tag customers, score their engagement, and set up page rules. As someone reads more, engages more, and downloads more, they get appended with new tags. This can trigger actions like adding them to a specific Slack channel or updating a newsletter segment, allowing me to give them a much better, tailored experience.
Moving forward, I plan to add new modules like membership areas, content protection, support desks, and dynamic content options. For instance, a regular visitor who engages more will have gated content automatically lifted, treating them to more valuable insights than a first-time visitor would see.
The Bottom Line
For me, vibe coding works out far cheaper than the traditional route of writing out a massive spec sheet, hiring a developer, and spending months going back and forth through chat trying to get it built. This way, I get the benefit of talking directly to the "developer" straight on my screen and fixing issues in real time.
I actually think this shift will ultimately make human developers stronger. The best developers will have the skillset, the structural knowledge, and a massive head start on how to leverage these same AI tools to form proper designs and scripts.
For now, I’m going to keep building on this journey with my small AI team: a project manager, a brilliant coder, and myself as the ideas guy who runs the smoke tests and judges the user interface. I look at the code strictly from the view of the customer, not the developer. To my mind, if it doesn’t feel right when you’re using it, we need to find a workaround to code it differently, rather than forcing the user to adapt to how it was built. I truly believe this delivers a better user experience, but as always, the proof will be in the pudding.
If you've tried vibe coding for yourself and built something, please drop a link in the comments form below! I'd love to take a look at what you’ve been working on. It’s amazing to see this new community of creators rising up and bringing what’s in their minds and hearts to life.
