We don't sell AI strategy.
We run the system.

Artificial intelligence is no longer the preserve of large corporations or digital start-ups. Across the private sector, French companies of every size are equipping themselves. In January 2026, France Num — the government portal for business digitisation — found that 32% of French companies expect to adopt AI within the next twelve months but have not yet chosen a partner. That isn't a competitiveness problem or an agility problem; it's a problem of choice. Between a consulting firm that sells a vision and an integrator that sells a generic tool, few players actually build and operate the system that has to run, day after day, on real data.

That gap is where we work — not by promising innovation in the abstract, but by building, one company at a time, the specific system most organisations are still missing. We keep our contribution deliberately narrow and deep, the opposite of programmes that promise to help every company with the same tool regardless of size.

Among the leaders this subject reaches, the first contact rarely comes from the same place. On a cash reconciliation, it's often a CFO who spots the bottleneck. On sorting inbound requests, it's a project lead or an operations manager — or the owner directly, in a leaner company. On a cross-cutting system, general management or IT makes the call. In every case, the assessment is the shared entry point — not a different offer depending on who reaches out.

What comes out of it often takes the form of dashboards the team can use directly — not one more tool to interpret, but an answer to the question they started with.

It isn't a matter of opinion: French law — the PACTE Act — sets the threshold, and INSEE applies it, sorting companies into four categories: micro and very small businesses, small and medium-sized enterprises (SMEs), mid-market companies, and large corporations. Three criteria classify each one — headcount, annual revenue, balance-sheet total. A mid-market company (in French, an entreprise de taille intermédiaire, or ETI) employs between 250 and 4,999 people, with revenue under €1.5bn or a balance sheet under €2bn; below that it's an SME, above both thresholds a large corporation.

The distribution makes the category striking: of the millions of companies in France, the overwhelming majority are micro and small businesses, and large corporations number in the hundreds — between them sit only about 4,400 mid-market companies. Few, but decisive: whether based in the Paris region or across the country, they carry a share of exports, employment and growth out of all proportion to their number. A mid-market company is neither a large SME nor a small multinational: it has its own decision rhythm — neither the endless purchasing cycle of a big group nor the tight budget of a very small firm.

The category itself is recent — created by the 2008 economic modernisation law and completed by the PACTE Act — and so is its ecosystem: Bpifrance, France's public investment bank, exists precisely to widen SMEs' and mid-market companies' access to finance (lending, accelerator programmes, dedicated savings vehicles), while its research arm and independent think tanks document a productive base too large for aid designed for small businesses, and too small to structurally interest the biggest firms.

Faced with a process to automate, a decision-maker really weighs three options. Hire — but the market for data and AI talent is tight, hiring is slow, and a single profile rarely covers the whole spectrum. Outsource to a firm — expertise arrives fast, but it leaves with the invoice, and dependence sets in. Or the third way, ours: build the system with you, then let it live on your side — your teams operate it day to day, and we stay in support to maintain it and make it evolve.

In other words: outsource what is built once, internalise what is used every day. That, to us, is the only split that doesn't create artificial dependence — in either direction.

We say it as plainly as the rest: a small business with no budget for AI isn't our target, and neither is a large group with a 12-to-18-month purchasing cycle. Big corporations and groups generally have their own in-house teams or framework agreements with major firms — not terrain where our format (a short assessment, a fixed-price sprint) makes sense. Plenty of agencies promise to serve companies of every size; we don't. Our target isn't an industry: it's a company size and a type of process. A fast-growing SME approaching the mid-market thresholds, on the other hand, is already part of the conversation — what matters, in the end, is the trajectory and the process, not the size of your company today.