AI companies love to be faceless. Here's a face.

I've spent twenty years building the web for businesses - first as creative director of a digital studio, then a long decade freelancing on builds for enterprise clients and award-winning agencies, with names like Toyota, Lexus, Deloitte and Yodel on the client list, and most recently leading the web engineering team at a UK digital agency: architecture, standards, CI/CD pipelines, code review, and the unglamorous discipline of keeping production systems alive for years at a time.

Somewhere in the middle of that, the tools changed. I started building with language models the way I once built with databases: AI agent pipelines, retrieval systems over business knowledge, AI-assisted engineering for code, planning, and quality control. Not because it was fashionable, but because it was the biggest change in what software can do for a business since the web itself - and because most of what was being sold around it was a demo wearing a suit.

Animus is the company I built to do it properly: connect the systems a business already runs, put AI workflows to work on the repetitive middle of the week, keep humans holding the pen on anything that matters, and stay accountable for the result - in writing, with a guarantee. The whole company runs on its own platform, and we published exactly how, mistakes included.

Before any of this I took a psychology degree, which turns out to be the least wasted part of the CV: automation succeeds or fails on whether the people using it trust it, and trust is designed, not assumed.