I spent four years training as an automotive engineer: first-class honours from Lancaster University, taught at Blackpool and the Fylde College. Four years of CAD models, CFD simulation and control algorithms, learning to design the machines that move people.
Then, a few weeks after graduating, I watched generative AI start doing the work I had just trained for. It was producing CAD drawings from a text prompt, setting up CFD simulations, and writing the control algorithms engineers used to tune by hand. The overlap between what AI could already do and what I had spent four years learning was bigger than anyone in the engineering department would say out loud.
That realisation changed what counted as the most useful thing I could build. If AI could handle some of the hardest technical work in engineering, the same shift was coming for the everyday work of every business: the quoting, the follow-up, the paperwork. Somebody had to build those systems properly, and very few engineers were choosing to.
Once I started looking, the pattern was everywhere: businesses paying senior salaries for clerical output, paying for twelve SaaS subscriptions that almost fit, and queueing every decision on one founder. None of it is a people problem.
It's all a systems problem.
So I build the systems.