AI did not change the process.
It removed the wait.

— The only thing that matters

What AI does. What I do.

I do not write code. I have never written a line of production JavaScript. But I have shipped 54 interactive tools, a 16-page portfolio website, and a PWA that improves daily task completion by 40%.

The division is simple: AI executes. I decide. AI generates HTML, CSS, and JavaScript in minutes. I decide whether the generated output solves the actual problem. AI fixes bugs when I point them out. I decide which bugs matter.

This is not "vibe coding" — throwing prompts at an LLM and hoping. This is requirements engineering at a higher level of abstraction. I am still doing BA/PO work. I am just doing it faster.

🤖

What AI Does

  • Writes HTML, CSS, JavaScript
  • Builds forms, dashboards, calculators
  • Makes changes when I describe them
  • Generates code in minutes
  • Fixes bugs I identify
  • Produces grammatically perfect copy
🧠

What I Do

  • Decide what features to build
  • Prioritise competing needs
  • Understand business strategy
  • Know when to say no
  • Test with real users and iterate
  • Make product decisions under uncertainty
Principles

How I make decisions

01. Ship or Kill. No Limbo.

Every prototype gets 5 days. Either it proves value or it dies. There is no "let me think about it" folder. Indecision is more expensive than a wrong decision because it consumes attention you could spend on the next problem.

02. Test with Real Humans. Not Theory.

I do not trust my own intuition about usability. I show the tool to 3-5 colleagues, say nothing, and watch. If they are confused, the design is wrong. Not them. The design. One fix. Ship again.

03. Free Means Free. No Gates.

No signup walls. No "create an account to export." No data collection. If a tool requires your email to function, it is not a tool — it is a lead generation funnel dressed as a tool. I do not build those.

04. Document the Kills. Not Just the Ships.

Most portfolios are highlight reels. That is marketing, not credibility. I show the ideas I killed and why. Because the fastest way to build trust is to prove you can recognise a bad idea before it costs someone $50,000.

05. Transparency Over Polish.

I show my $60/mo stack. I show my failed prototypes. I show the exact prompts I use. Not because I am generous — because hiding the process implies the process is the value. It is not. The decisions are the value.

06. Tools Solve Problems. Not Categories.

I do not build "a stakeholder management tool." I build "the thing that stops 73 banking executives from sending conflicting emails during a merger." Specific problems. Specific tools. Generic solutions help no one.

The Stack

Full transparency on what powers this site and every tool. ~$60/month. 1,100x ROI vs traditional agency development.

See the Full Stack