The AI boundary luxury brands keep getting wrong

Luxury is adopting generative AI faster than it is defining where the technology truly belongs. Deloitte’s Global Powers of Luxury 2026 report, which surveyed 420 senior executives across 10 countries, found GenAI shifting from exploration to execution, with 41% of companies deploying it in selected areas and nearly 67% of executives expecting revenues to hold or rise this year. What the adoption data does not capture is the difference between deployments that support a brand’s cultural position and those that quietly dismantle it. That distinction is becoming more urgent as a second shift unfolds outside the organization. McKinsey’s directional survey of luxury consumers and merchants, published in May, found 85% of luxury consumers already use a multipurpose AI assistant such as Gemini or Perplexity to support shopping decisions, with 52% doing so frequently. Desire formation, comparison, and the first interpretation of intent are moving upstream into interfaces brands do not control. How houses answer both questions — where AI belongs internally and how they are represented externally when an agent speaks first — will determine whether they scale with a distinct point of view or toward greater sameness. The hidden cost of luxury’s AI adoption Efficiency metrics measure what AI saves. They do not measure what it erodes. People pay a luxury price for human input — the eye that made the selection, the judgment that decided this and not something else. When that input disappears, the damage does not surface in a single quarter. It accumulates across every piece of content that carries no identifiable human decision behind it, degrading a cultural signal built over decades, slowly enough that organizations rarely connect cause to effect until the distance is difficult to close. Audiences register the absence before they can name it. “When an AI-generated campaign gets criticized, it is not really about the tool. People can feel a lack of authorship behind it,” says Margarita Repina, founder of Atelier Catalyst. GenAI: Enhancement or replacement? The real question is not whether a function should use AI, but where AI enhances the work and where it begins to replace the judgment that makes it valuable. Kering chief executive Luca de Meo framed the ambition at the group’s Innovation Day in Paris on January 22, asking how AI can help creative people get “better, faster, stronger” rather than replace the creative process. When that sequencing holds, the technology accelerates creative work; when it collapses, it accelerates convergence. A 2024 study by Anil Doshi and Oliver Hauser in Science Advances found writers using AI assistance produced more creative work individually, but their outputs became significantly more similar overall. GenAI optimizes toward the mean, and the mean is the enemy of distinction. Operations show the clearest gains at the lowest cultural cost. LVMH’s partnership with Google Cloud, effective since 2021, has built centralized data infrastructure across its 75 maisons for demand forecasting, inventory, and pricing. Outside luxury, Zalando said last year that GenAI cut its marketing image production from six to eight weeks down to three to four days at 90% lower cost. When aesthetic direction is grounded in genuine judgment and AI supports execution, the brand signal remains intact. Communications is where the boundary has already been tested in public. Hermès relaunched its website in January with 12 hand-drawn illustrations by French artist Linda Merad — the paper grain left visible on screen and the line deliberately imperfect — and drew praise across the industry. One month later, Gucci posted AI-generated teaser imagery, labeled “created with AI,” ahead of Demna’s debut Milan show, and critics dismissed it as “AI slop.” Two decisions, two positions in the industry’s cultural hierarchy. Judgment, in the end, is the input that determines output quality. A prompt specifying the quality of light or the texture of a surface comes from someone who has spent years looking carefully at things, not from someone who learned to operate a tool. De Meo made the investment case at the same Innovation Day: carmakers devote up to 10% of turnover to research and development, while the luxury sector invests around 1%. Closing that gap means hiring people with a trained eye to direct the tools — what we would call creative technologists — rather than buying more tools. When AI agents shape luxury demand The same question of balance extends beyond the organization. As AI reshapes how brands operate internally, it is also changing how consumers discover them externally. According to Liya Dashkina, co-founder of FashionTech Middle East, “Luxury brands will need to optimize their infrastructure for AI discoverability in the background while preserving the human magic of discovery in the foreground.” Agents optimize for what they can measure: price, availability, popularity, and structured product data. They do not optimize for meaning, restraint, or cultural codes. McKinsey identifies the risk as the “flattened” brand: the agent does not get the SKU wrong but frames the brand incorrectly, translating a house defined by discretion into a set of comparable attributes. Its survey found that luxury consumers are almost evenly split on where they expect to interact with AI in the future, including search engines, personal AI agents, and brand-owned apps. This means no single channel can fully reclaim the first interpretation of desire. Commercial infrastructure is arriving faster than most luxury operating models can absorb. Swap launched what it calls the first agentic storefront on May 12, claiming conversion at twice the industry rate; Amazon began licensing its agentic shopping technology to retailers through AWS on May 27, citing conversational sessions that convert at 3.5 times the rate of keyword search. Both figures are the companies’ own. Meanwhile, the EU’s textile Digital Product Passport rules — with a delegated act expected in 2027 and compliance beginning around 2028 — will push brands to make provenance machine-readable, creating an opportunity to encode cultural codes directly into how agents represent them. Trust in this environment is built before the consumer ever opens an app. “Brands have to earn their place in the agent’s model of the consumer before the consumer even opens the app,” Anne-Liese Prem, head of cultural insights and trends at creative agency Loop, says. Her test for content is an authorship filter: what a brand has refused in the past year is as legible as what it has produced. Takeaway for luxury leaders Luxury consumers in McKinsey’s survey estimate 39% of their purchases will be agent-mediated by 2030; merchants expect 47% of interactions. The share matters less than who governs the interpretation of intent as mediation scales. Scale with authorship compounds recognition. Scale without it produces content agents cannot trust, consumers cannot place, and culture cannot remember. The brands getting this right are not slowing their AI adoption — they are applying it with greater precision around the elements that make them irreplaceable.

The AI boundary luxury brands keep getting wrong

How Bottega Veneta weaved China’s forever it bag

Upgrade to Pro

Sectors

Insights

Company

The AI boundary luxury brands keep getting wrong

How Bottega Veneta weaved China’s forever it bag