Bringing artificial intelligence closer to the real world.

Conscientia

HOMN starts with a fintech. With Conscientia, we wanted to put artificial intelligence at the service of finance: helping people step back before investing, using precise and often personal data.

Very soon, one question stopped us. For an AI to work on this data, it had to be entrusted to a model provider, on servers we did not control. For a company that talks about money and trust, that was not acceptable.

So we looked for another way: open models, run on machines we control. It worked. And it changed everything about how we saw AI.

The beehives

At the same time, Lucas was working with an association that wanted to protect beehives from the Asian hornet. It had to be spotted before it attacked the bees.

Lucas designed an appliance with an intelligence able to tell a hornet from a bee at the hive entrance. In a field there is no fibre and no datacentre. The decision had to be made on the spot, in a fraction of a second, without asking a distant server for permission.

The turning point

On one side, data that must not leave. On the other, terrain the network does not reach. Two problems that seemed unrelated had the same answer: put the intelligence where the data and the people are.

Talking to those around us, we realised this need was everywhere. A law firm protecting its clients’ secrets. A club that wants to keep its players’ medical data. A factory that wants to modernise a machine without replacing everything. A local authority handling its residents’ files. They all wanted the same thing: an AI of their own, on their premises.

The risk

Today, the most powerful models are either closed and American, or open and Chinese. Europe uses both, but controls neither the first nor the infrastructure that runs them.

A political decision can cut off access. An outage at one provider can stop thousands of companies at once. A change of price or terms takes effect overnight. When AI becomes an everyday working tool, depending on it entirely becomes a risk to business continuity.

The choice

We are not against the cloud. The best models live there, and it is often legitimate to use them. What we want is for it to be a choice, data by data, task by task.

A HOMN appliance works on your premises first. It connects when you decide. It accepts any model, and you switch when a better one arrives, without rebuilding anything.

The meter

We believe AI will become a basic resource, like electricity. No one wonders where the current comes from when they switch on a light. They rely on a meter, at home, that works.

That is what we are building: every company’s intelligence meter. A machine set up on your premises, that gives access to the best models, keeps your data, and keeps running when everything else stops.