You ask an AI assistant for a statement bag. It returns three options. All three are owned by the same conglomerate.
That is not a neutral recommendation. It is a distribution strategy.
Kering launched KNXT last month. Madeline is their AI shopping assistant, and it draws from a portfolio that spans Gucci, Saint Laurent, Bottega Veneta and 27 other brands. When a customer asks Madeline for a leather bag under €2,000, it recommends a Kering brand. When they want a fragrance, Madeline recommends a Kering brand. When they want ready-to-wear, Madeline recommends a Kering brand.
This is how the portfolio conglomerate model adapts to the agent era. They own the brands and they build the agent. The agent stays inside the ecosystem by design.
The scale of what is happening in agentic commerce is worth stating plainly. 45% of consumers now use AI for some part of their buying journey. Brands using agentic commerce are seeing 3x conversions and 38% average order value uplift. These are early numbers from early deployments. They will grow.
The question is not whether AI agents change luxury distribution. That question is settled.
The question is who controls the recommendation layer when they do.
Kering controls it for Kering. LVMH will control it for LVMH. Richemont will control it for Richemont. The portfolio conglomerates will build agents that serve their portfolios. Even a genuinely helpful assistant built inside a group cannot recommend a Celine competitor. The architecture makes that structurally impossible.
Independent maisons cannot build their own Madeline. The economics do not work. A single brand cannot produce a cross-portfolio AI that a customer finds compelling enough to use.
So what do independent brands do?
They need to appear on someone else's agent. On the ChatGPT shopping layer. On Perplexity's product search. On whatever Tata CLiQ Luxury builds next for its Vue.ai integration. On the general-purpose assistants that consumers already use before they ever open a brand website.
Getting onto those lists is not a marketing problem. It is a data problem.
AI agents select products based on materials, durability, sizing, price-points. They query structured data. They do not query heritage narratives. They cannot. None of that is machine-readable.
I wrote about this last week, and the Kering announcement makes the argument more specific. Before, the threat was abstract: AI agents prefer structured data, and luxury has not built it. Now there is a named agent, a named portfolio, and a named winner by default.
The brands that structured their product catalogues in 2025 will be in the datasets that third-party agents draw from. Machine-readable provenance. Accurate sizing schemas across markets. API-accessible pricing and availability. The brands that did not will not appear on a shortlist they cannot see being made.
A few names worth watching.
Tata CLiQ Luxury is building AI-native luxury infrastructure at the platform level, not the brand level. Brands on the platform inherit some of it. The ones that are not, do not.
Swiggy has live MCP integration in India. Flipkart is in development. Model Context Protocol is the emerging standard for how AI agents connect to external product data. The first independent luxury brand that publishes an MCP-accessible product catalogue does not just get a press story. It gets agent distribution across every assistant that follows the protocol.
Is that guaranteed to matter by 2027? Genuinely unclear. But so was mobile commerce in 2010. (The brands that dismissed it then are still catching up.)
I want to be clear about what I am arguing and what I am not.
Heritage still matters. The boutique still matters. The craftsmanship story still matters. What is changing is the discovery layer. Not making product data machine-readable does not preserve a brand's intangibility. It just makes the brand invisible to a channel that now handles 45% of the buying journey.
The question I am sitting with: who is the first independent maison to get onto a third-party agent's recommendation list, and what specifically did they do to get there?
Not Gucci. Not Saint Laurent. They already have Madeline.
Someone else. If you are working on this, I would genuinely like to hear from you.
Sources: Alhena AI 2026, BoF State of Fashion 2026, W19/W20 Research Brief, Tata CLiQ Luxury via Vue.ai and Business Standard.
