I built 3 AI products in 3 weeks. Here's what nobody tells you about building with AI.

Three weeks ago I had no products, zero revenue, and absolutely no idea how to vibe code.

Today I have two deployed SaaS apps and a full AI automation system that books appointments for businesses. This is what that actually looked like.

A compliance dashboard for UK/EU green claims regulations. I built it with Whop, Claude Code, Codex, and Cursor. It took about a week.

I found the niche using reverse keyword searches and Google Trends to identify a high volume, high margin, low competition intersection. That research process alone is something I'll keep using.

A local SEO and AI visibility reporting tool for agencies. I built this one with Supabase, Clerk, Vercel, Claude Code, Codex, and Cursor.

This was a lot more complex, but it only took four days. The first build forced me to learn the tools... so the friction from week one bought me speed in week two.

An AI lead conversion system built on n8n, OpenAI, Twilio, and Airtable. The first version took me about 4 to 6 hours.

I finally got a test text message working where the system took over entirely - the AI replied, qualified the lead, and pushed it straight to the CRM. I built it as a reusable template, so deploying it for a new client now takes 15 minutes.

Here's what nobody on Twitter tells you when they post their £10k MRR screenshots.

Vibe coding sounds easy. But when something breaks... you still have to debug it. You need to understand enough about the backend to actually fix the mess.

You aren't just using one tool. You're combining Cursor, Claude Code, Codex, databases, APIs, and hosting platforms. It's a stack, not a shortcut. Learning how these pieces talk to each other is the actual work.

Distribution.

I can build a product in a week. But getting the right people to find it, use it, and actually pay for it... that's a completely different game.

And it's the one I'm playing right now. I don't have it all figured out, but I'm solving this problem in public, and I'll share everything I learn along the way.

In future emails, I'll share the honest truth about my distribution tests - what's working, what's bombing, and the templates I'm using to try and get these tools into the hands of paying users.

I'll also be sharing highly practical AI tools and systems that actually make money, save time, and add real value to your day.

If you're building or just want to see how this stuff actually works behind the scenes, subscribe so you don't miss the next one.

You can also connect with me on LinkedIn or Twitter - I post shorter versions of this stuff daily.