I was designing AI products in 2017, before it was a title. My first one raised its next round a month after I shipped it.

The surface keeps changing. The question never does: how should this AI behave, and when should it get out of the way?

At Rocket Lawyer I lead Copilot. I turned a pile of features into one system that moves people through a legal matter, and changed what the business sells in the process. I also dragged the way I work, vibe-coding, fast prototypes, into a 16-year-old company that wasn't built for it.

I'm best before the roadmap exists, when nobody's sure what the product should be yet.