I spent years inside other people's businesses before I started
automating them.

I started out as a stockbroker, working with numbers, risk, and money that belonged to someone else. That kind of work builds a specific habit: check it, then check it again, because being roughly right isn't good enough. That habit is why I test systems before I trust them.

When ChatGPT landed I went deep on the creative side - images, logos, music. What stuck wasn't the output. It was watching the same tool produce completely different results depending on how I asked. Being specific about what I wanted changed everything. That's still true - most of what makes a system work well is being clear about the job before anything gets built.

I moved into fractional executive assistant work, which put me inside the day-to-day of very different businesses - a fractional CFO, an education consultant, real estate, other consultants, healthcare professionals. I saw how work actually moves - where it stalls, what gets dropped, and how often a "process" is really just one person remembering. I didn't have a name for it then. It was systems thinking.

I started building the systems myself - dashboards, task systems, websites, automations - and then got serious about the fundamentals underneath them. Evals, error handling, logging, knowing what breaks. An automation that works today isn't the same as one that still works in six months. Building for the second one is the actual work.