About Zonai
Amsterdam-based. Production-obsessed. Est. 2025.
Most AI projects fail between the demo and the deploy. We built Zonai to fix that.
We started because we kept seeing the same pattern: impressive proof-of-concepts that couldn't handle real users, real languages, real edge cases. Our entire company is organised around one thing: getting AI systems into production and keeping them there.
Started with a single question.
Two people and one obsession: why do AI projects look great in demos and fall apart in production. We decided to find out the hard way. By shipping real systems for real users.
We learned by doing the hard stuff.
Our first production system was a multilingual moderation pipeline. Half a million messages a day, no tooling, a small team overwhelmed. We built it in six weeks. It's still running. Then we grew. Engineers. A designer. A linguist who kept telling us our models didn't understand how people talk.
We stopped saying yes to everything.
We turned down projects that didn't fit. We picked three areas we could go deep on. No more buzzword menus. We added a research function and started building the kind of team that argues about whether a model is ready to ship.
Still shipping. Every day.
We work with companies across regions and industries. The question we started with hasn't changed. We're better at answering it. Every project has a lead who owns it end to end and loses sleep over it.
Things we believe.
Not values from a workshop. Things we've learned by getting it wrong first.
Ship or it didn't happen.
A prototype that never goes to production is a very expensive hypothesis. We measure ourselves by what's running. Not what's in a deck.
Honesty over optimism.
AI is hard. We tell clients when something won't work. When a timeline is unrealistic. When the data isn't good enough yet.
Small teams, deep craft.
We've never grown faster than our ability to stay good. Every engagement has a lead who owns it completely. Not a rotating cast.
Practical, curious everywhere.
We believe in building things that work. We also collaborate with researchers and engineers across many countries and contexts.
Come find us in Amsterdam.
We're always happy to talk about a project, a problem you're stuck on, or what's happening in applied AI right now.