Gonzague de Pirey joined LVMH in November 2022, the same month ChatGPT launched. Three years later, at VivaTech, he is the one naming AI's edges inside the world's largest luxury house. VivaTech gathers Europe's biggest tech names every year; choosing that stage to talk about boundaries, not numbers, is already a communications decision.

Seventeen days earlier, Marta Ortega addressed Inditex shareholders for the fifth time since 2022. Same industry, opposite ends: a luxury conglomerate and a fast-fashion giant, both speaking about AI within two and a half weeks of each other. Inditex is, by sales, the largest fashion group in the world. Ortega acknowledged AI will transform much of the industry, then closed the point with one line: "the true value is in people." No system named. No boundary drawn. The contrast in register tells its own story: one company is treating AI as an operational question, the other as a values statement.

A shareholder-meeting line is never written for one listener.

LVMH calls its position "quiet tech": AI deployed across the value chain, but explicitly at the service of craftsmanship, traceability, and the human relationship with the client. De Pirey has held that line since 2025. He names the domains where AI operates. He names, too, what stays outside them.

Ortega hasn't done the same — and probably didn't need to. A statement like that is never written for a single audience. Shareholders hear risk contained. The market hears no number that moves anything. The media hears the headline. Employees hear — or fail to hear — the only question that actually affects them: what changes in their job on Monday. That's not a flaw in Ortega's remarks — it's what happens when one line has to satisfy four audiences that never asked the same question.

LVMH

"Quiet tech": boundary named

Inditex

"People": no boundary

The test

Verifiable in 12 months

Specificity isn't a style choice. It's what turns a values claim into a claim you can check in twelve months.

Next time a leadership statement on AI sounds reassuring, it's worth asking one question: does it name a boundary you could verify later, or only a value you already agreed with?