Builder, advisor, and translator between AI strategy and shipped systems.
I help organisations decide what AI should do, build the first useful version with their own people, and teach the team to operate it, without turning every decision into a vendor dependency.
The work is deliberately practical: fewer slogans, more workflows, clearer interfaces, stronger judgment about where agents belong. I work best where executive ambition meets operational reality.
Three principles that survive contact with operations.
- 01
Start with the workflow
The model is never the strategy. I look for the repeated decisions, hand-offs, approvals and information gaps where AI removes real drag, and ignore the rest.
- 02
Make the system legible
Useful AI has visible inputs, clear stop points, and outputs a team can inspect. Black-box magic does not survive contact with operations.
- 03
Leave ownership behind
The goal is not dependency. Simple interfaces, documented patterns, and enough teaching that your team keeps moving after I have gone.
Where the work has taken me.
Two AI products of my own, a company brain built for Zazi, and AI strategy and delivery inside large organisations. The products keep me honest about shipping; the advisory about what operations can absorb.
- HeathrowBuy-and-build strategy and tender guidance
- RabobankIT project management for the mortgage application system
- Shell
- CZSet up the low-code platform that enabled development
- Valcon
- ING
- Zazi
- The advisory workWhat each engagement involved
Why it looks the way it does.
This site borrows its palette and its dappled light from late-1870s Impressionism: ivory ground, a little rose, sage and sky, and colour placed next to colour rather than outlined. I chose it because that way of painting prizes attention over spectacle: ordinary afternoons, looked at properly. That is roughly what good work in my field looks like too.
Bring the messy version. We’ll find the useful system inside it.
Replies within one working day. Or email yvoderooij@gmail.com.