The AI-Native Workflow: From Problem Statement to Working App
Building apps used to mean weeks of writing PRDs, endless wireframes, and months of waiting for a working prototype. In an AI-native world, that cycle looks very different.
Recently, I set out to build a lightweight app idea — and instead of starting with a blank page, I started with AI. From drafting and refining the PRD with ChatPRD to generating a working prototype in Lovable, the process felt less like traditional product development and more like a fast-paced conversation. Here’s how it unfolded.
🔍 The Problem I Wanted to Solve
Every app starts with a problem worth solving. For me, it was:
👉 Most of us are in WhatsApp groups for sports like 5-a-side football. But organising a game quickly spirals into chaos: overlapping replies, last-minute changes, and nobody sure who’s in or out. Afterwards, reconciling payments is even worse — someone always ends up chasing money or juggling spreadsheets.
Rather than describe it in text, I recorded a short video to explain the “why” behind this app:
With the problem nailed, it was time to turn ideas into something concrete.
📝 Turning an Idea into a PRD with ChatPRD
With the problem clear, I needed a Product Requirements Document (PRD). Normally, this step feels heavy — multiple drafts, lots of stakeholder back-and-forth.
This time, I treated AI like a collaborator. Using ChatPRD, I:
- Drafted the first version in minutes.
- Iterated by asking “what if” questions and testing alternative approaches.
- Tightened the scope until the PRD felt both practical and inspiring.
What surprised me most: this step became less about “writing a document” and more about having a conversation with the product itself. Instead of me laboring over structure, AI helped me explore possibilities quickly.
⚡From PRD to Prototype with Lovable
Once I had the PRD, the natural next step was building. This is where Lovable came in. I dropped the AI-refined PRD into Lovable and, in a short span of time, had a working prototype.
No lengthy dev cycles. No static wireframes. Just something real that I could click, test, and share.
Here’s a quick look at the demo app in action:
💡What “AI-Native” Means to Me
For me, building in an AI-native world isn’t about replacing product skills — it’s about amplifying them.
- Speed: Move from idea to prototype in days, not months.
- Iteration: Explore multiple paths quickly before committing.
- Focus: Spend more time on vision and decisions, less on admin.
Would I do things differently next time? Absolutely — AI still has limits, human judgment is non-negotiable, larger team and projects still need to be grounded in sound engineering practices. But the shift is clear: the way we build apps is changing, and it’s hard to imagine going back.
I’d love to hear how others are experimenting with AI-native tools. Have you tried building this way yet? 🚀
Originally published on LinkedIn.