Decisions made under pressure determine outcomes wherever the alternative is finding out afterwards; A pricing change that ends a customer relationship, a triage protocol under load, a crisis response published before anyone knows how it will land, or a mission plan where hesitation is measured in lives. The domains differ, but the structure does not. Our work is to give people clarity on consequences before the decision is made rather than after it is irreversible. This is not theoretical, as every improved decision in a high-consequence environment has a direct human cost attached to getting it right.
Europe should not have to rent its decision infrastructure from somewhere else. The behavioural record a model trains on is not a neutral commodity: whoever holds it decides who may ask questions of it, on what terms, and for how long. Ours is held in and for the benefit of Europe. Client data stays governed by the client rather than absorbed into a model that goes on to serve someone else's customers. That matters to a hospital group and a games publisher exactly as much as it matters to a ministry. Strategic autonomy is really a question about who owns the record of how people behave, and Europe is currently answering it by default.
We started in competitive gaming and digital communities because that is where behaviour is measured most completely. Millions of people cooperating, defecting, signalling status, forming groups and leaving them, all of it recorded continuously by products they were already using, none of it performed for a researcher. Nobody set out to build a behavioural instrument, but one accumulated anyway. What it captures is not gaming-specific: cooperation, trust, defection and belonging behave recognisably wherever people are put under pressure together, which is why the same substrate informs decisions far outside the industry it came from.
~2000 BCE - 1800S
Games as training instruments. Chess taught generals encirclement. Go modelled territory control. Mahjong taught probability under incomplete information. Simple rules, profound strategic depth.
Thirty moments where structured play, formal method and machine learning arrived at the same question: what will a group of people do. Drag to move through them.
PATTERN RECOGNITION
None of this is new. Structured play has been a laboratory for consequential decisions for four thousand years, and formal method caught up with it a century ago. What changed recently is only the resolution of the record and the cost of running it: a population that would once have been a thought experiment can now be run a thousand times before lunch.