Because buyers deserve better than generic tools: an approach built by the craft, held to its rules, deployed with the confidentiality that prices and suppliers demand. Yxalo was born in the field — not in an agency.
Twenty years in industrial procurement — the plant, the supplier crises, the negotiations, leading procurement teams in France and across Europe. Then four years training buyers. And everywhere, the same observation: the craft spends an enormous share of its time on preparatory work — reviewing, compiling, re-reading, re-drafting — at the expense of the gesture that creates value.
In parallel, a market observation with no ambiguity: buyers already use AI, massively — but on consumer tools, with no procurement specialisation, no craft perspective, no operating method. The decisive step was missing: professionalising AI for the craft.
Yxalo was born where those two worlds meet: the knowledge of industry and procurement on one side, the demand for confidentiality and teaching on the other. Assistants designed by people of the craft, within a frame that respects each company's culture and DNA. Nobody offered that to buyers. So we built it. Craft first, AI inside.
20 years of experience in automotive purchasing and supply chain, up to regional procurement leadership in Europe. He notably led the deployment of industrial sites in Mexico and China, as well as commodities on a global scale.
A procurement assistant is not judged on a demonstration. It is judged on your real files: RFQs, 8Ds, contracts. It must produce reliable, explainable analyses and clearly flag what it does not know. That is how we build the Yxalo assistants. — Nicolas Mattiuzzi, founder
Yxalo mobilises experienced professionals in procurement, supplier quality and supply chain, selected according to the sector and the file — across France and Europe, in French and English.
A proprietary method of structuring, verification and craft validation, applied by procurement professionals.
AI prepares, the buyer decides: every figure gets verified, every recommendation gets arbitrated. AI never invents — and never decides.
The first demonstration runs on fictitious, public or previously anonymised data. Your real files are processed only after the contractual framework is in place and the appropriate confidentiality level has been chosen.
