The Agentic Fashion Report

The agent layer is where the gap between intent and infrastructure becomes most visible. LVMH describes 'agentic automation' as being 'still in its infancy' inside the group, even as it runs MaIA at scale and deploys what Soumia Hadjali of Louis Vuitton calls an 'intelligent layer to superpower our client advisor.' That framing — agentic as a client advisor amplifier, not a transaction engine — is worth holding. The external agentic commerce protocols gaining traction in 2026 (Google's Universal Commerce Protocol, OpenAI's Agentic Commerce Protocol via Stripe) are designed for e-commerce checkout flows, not for relationship-intensive luxury clienteling. The luxury brands making progress on the agent layer are not plugging into these protocols. They are building proprietary stacks: LVMH's MaIA, Brunello Cucinelli's Callimacus multi-agent site composition, Burberry's Penguin generative discovery and recommendation platform deployed across 100+ global consultants with a 24% uplift in average transaction value.

What these initiatives share is that they are advisor-facing, not consumer-facing. The agent is inside the boutique, helping the human who serves the client — not replacing the human in the conversation. This is not conservatism. It is a well-reasoned thesis about where AI generates compounding returns in luxury: when the advisor has better context, the client relationship deepens, lifetime value rises, and the brand accrues data that refines future recommendations. The agent layer for luxury is a clienteling infrastructure layer, and the brands building it are building a durable competitive asset. The brands that have not started — and the Protocol Readiness dimension makes clear that most tracked brands score in the low 20s here — will find that protocols alone do not create the data foundation the advisor layer requires.