I Shipped My First App to the App Store and What it Means for Product Development 🚀

As long as I’ve been a PM, I’ve had more ideas than engineers than engineers to build them.

As long as I’ve been a PM, I’ve had more ideas than engineers than engineers to build them. Much of what inspired me to build was marooned in my mind or in paper notebooks sitting on my bookshelf. Even for someone who works primarily with engineers, it has not been easy to make side projects come to life.

That started to change last year, and last week, I did something I never expected: I shipped an app on the Apple App Store! 🎉

Vibe Coding: The Early Year (2025)

About a year ago, I went deep comparing AI tools for anything I could throw at them; I tried ChatGPT, Gemini, NotebookLM and even tinkered with xAI. But I kept coming back to Claude. Something about its voice and writing style resonated with the way I think about problems. Then, I discovered tools that changed everything: Lovable, Cursor, and eventually Claude Code.

I started with Lovable and Cursor, vibe coding my first few websites. I built a few cool prototypes, including one I’m still a little heartbroken about: a PM interview grading app called Robin that used voice diarization to differentiate interviewer and candidate speakers. It actually worked.

Unfortunately, I switched laptops and I didn’t back it up and showcase it on my (now active) GitHub page. Lesson learned. 😅

Then I hit a wall. I tried to deploy a To Do list app to the App Store. Yes, I know everyone has had this idea. But I couldn’t get it across the finish line with my existing setup (Cursor and Xcode).

Something wasn’t clicking.

Going to the Source

Since Claude already powers both Cursor and Lovable under the hood, I asked myself: why not go directly to the source?

I cancelled my Cursor and Lovable subscriptions and upgraded my Anthropic sub to the Max Plan. One of the things I love about Claude Code is that I don’t need to fuss with an IDE for most tasks, and I have options: web, desktop, and CLI.

I was hooked immediately. I’d already saved $336 on hosting by using Claude Code to port my personal website from Wix to Cloudflare Pages (free and blazing fast). The ROI was proving itself. 💰

The Challenge: Ship Something I’d Actually Use Without Human Help

I gave myself a constraint: build an app I would personally use every day, and ship it to the App Store.

The idea for Now Meditation came from a book I read years ago called The 8 Minute Meditation that kickstarted my meditation practice (a habit I’ve mostly maintained since). The premise is simple: eight minutes of meditation per day is all you need to start a lasting practice.

Here’s the thing about meditation: you’re not supposed to worry about how much time you’re spending. That’s where a timer comes in. Sure, I could use the timer app on my iPhone, but it isn’t designed for mindfulness, it doesn’t help me track my practice over time, and I don’t get mindfulness minutes as credit.

I looked at what was out there. Most meditation apps are either expensive, cluttered with influencer content I don’t need, or both. I wanted something minimal, private, and focused. 🧘

What I Built

Now Meditation is intentionally simple, minimal, private and free.

• Clean a meditation timer. No guided sessions, no ambient sounds, no subscription upsells.

• Private everything stored locally on-device. No data leaves your phone.

• Optional Apple Health integration to track mindfulness minutes.

The Surprisingly Easy (and Hard) Parts

Building the app with Claude Code? Surprisingly easy. I jammed with Claude creating markdown-based input instructions for Claude Code. The AI handled the native Swift code, the UI layout, and the logic with impressive fluency. I could describe what I wanted in natural language and iterate fast.

The hard parts had nothing to do with AI:

• Xcode Apple’s IDE is powerful but dense. Navigating it as a first-timer was a learning curve unto itself.

• App Store Connect The submission process, provisioning profiles, certificates, and review guidelines were also difficult, but I figured it out.

My first submission was rejected. Apple was pretty good about reviewing Now quickly and providing actionable feedback on why they rejected it. I worked through the issues Apple flagged with Claude Code, then I resubmitted, and got approved. ✅

What This Means for PMs

The barrier between think and do is vanishing before our eyes. If you’re a PM with strong opinions about what should exist in the world, just build it. The cost of experimentation has collapsed and the feedback loop between idea and artifact is nearly instant.

I think this means neither PRDs (the traditional PM artifact), nor traditional product development teams (Design, Product, Eng) go away. But I do think not only based on what I’ve seen, but based on what I’ve been able to produce that product development is on the precipice of a dramatic overhaul.

If PM, or design for that matter, can create and prototype powerful, high-fidelity applications in hours without engineering support, we can accelerate user feedback, iteration, and literally anyone with a good idea can be the next founder of a unicorn company.

Does that mean engineering goes away? Absolutely not. I still believe these applications need careful and expert curation to ensure they’re safe, efficient and scalable for subsequent iterations.

What’s Next 🔜

The app is functional and needs some polish, as I focused on the problem of deployment over app-polish. I focused on deployment because any app I build will require this step. I’m continuing to iterate on the app, adding streaks and exploring what “just enough” gamification looks like to use technology to create a healthy habit.

If you’re a PM sitting on an idea, my advice is simple: pick the smallest version of it, open Claude Code, and start building. You might surprise yourself.

Have you tried building with AI tools? I’d love to hear what you’re working on, drop a comment or DM me. 💬