ORIGIN
It all started with a law firm.
About forty people, a managing partner convinced it was time to adopt AI, and a request that had stalled. Appliance was born from that conversation.
Six months waiting on Copilot licenses
We happened to meet the firm’s managing partner during a conversation about AI. For more than six months, he had been asking his IT provider for Microsoft Copilot licenses. The same provider that set up his computers, network and printers, the one he called the moment something stopped working. On AI, though, nothing was moving.
In the meantime, another offer had reached him: a solution designed for regulated professions, billed at around €300 per employee per month. For the firm, that came to €12,000 a month, or €144,000 a year.
He was ready to sign. Not because the price seemed fair, but because staying on the sidelines seemed costlier still.
Caught between two risks
Around him, AI was already everywhere. His peers were using ChatGPT or Claude, sometimes with approval, often without. The time savings were real, and sooner or later someone would paste a contract or a client file into a tool they didn’t control. Everyone knew it, and no one wanted to be the first firm blamed.
He found himself facing two options that worried him equally: wait and fall behind, or let it happen without knowing where the firm’s data was going.
As we listened, we realized his need lay elsewhere. He wasn’t looking to choose a model or to run a technical project. He wanted his teams to have a powerful AI, one he trusted and could roll out to the whole firm.
He didn’t want an AI project. He wanted AI.
The Wi-Fi router detour
The answer came to us from a far more ordinary object. Behind an internet connection sits a huge stack of hardware, protocols and security. Entire teams have made it their job.
Yet nobody thinks about it. You call your internet provider, a Wi-Fi router shows up, you plug it in, and the internet works. The complexity hasn’t disappeared: someone else takes care of it.
Why couldn’t AI for business work the same way?
All the complexity inside. A simple product outside.
A small box, sitting on a desk
That’s what we offered him. A small box installed on his premises: you plug it in, you set up access for his team, and the firm has its own AI. An app to talk to it, hand it documents, launch agents or do research. The data and the compute stay at the firm. What happens inside is our job.
The shift was visible right away. AI was no longer an abstract topic but an object, something you install, buy, pay off over time and hand over to your team. That’s why the product is called Appliance.
The same need, everywhere
Since then, we’ve run into the same situation at almost every meeting. An accounting firm, a manufacturer that keeps thirty years of know-how in its files, an engineering firm, a family-owned business. All of them have sensitive data and business constraints. None of them has a team dedicated to AI, and none wants to build one.
That’s who we build Appliance for. And the more we master what’s inside the box, the simpler it can stay on the outside.