An applied AI company,
built in Europe.
We build AI systems that companies run on. They do real work every day, on our clients' own data, inside the tools their teams already open every morning.
Getting a model is easy now.
Getting one into production is not.
Most companies have already run the pilot. It worked, everyone agreed it was impressive, and then it stopped there. What stands between a demo and a system somebody uses on a Monday is the plumbing, the evaluations, the compliance work and the hundred decisions about how the output reaches the person who needs it.
That part is unglamorous, it is most of the work, and it is the reason so much of this ends in a slide rather than a system. It is also the only thing we do.
Four things we hold to
Each of these survives being written into a contract, which is the test for whether it belongs on this page at all.
We sell outcomes, not hours
The scope, the baseline it has to beat and what finished looks like are agreed before anything starts. A project billed by the day has no reason to end, so we do not price one.
Production is the finish line
Nothing counts as delivered while it is a notebook, a demo or a deck. The engagement ends when the system is running where the work actually happens.
Decision support, never a verdict
Every output carries the evidence behind it and how far it should be trusted. A person decides what happens next, and a reviewer can check any single result rather than trusting the aggregate.
Your constraints are the brief
We do not ask for data, we do not ask for access, and we do not ask for a change in workflow. The system has to work inside the constraints of the client, and if it cannot, we say so.
If AI is not the gap, we say so and there is no project. That outcome is more common than the sales version of this page would suggest, and saying it early is cheaper for everyone than finding out in month three.
Where we are, and what we are held to
- Founded
- 2024
- Based in
- Barcelos, Portugal
- Compliance
- GDPR and EU AI Act, externally validated
- Deployment
- Our cloud and AI models, your public cloud account, or on-premise
Compliance was the starting point. The validation was done by an outside party, and it covers all three deployment models, including the two where the data never touches infrastructure we control.
Tell us the work you want to fix
One call, no deck. We will tell you whether this is something we can build for you, and if it is not, we will say that too.