Agentic AI Is Coming for the Customer Relationship — What Luxury Brands Do Now.

Shopping-related AI searches grew 4,700% between 2024 and 2025. ~80% of luxury consumers now use AI tools for product discovery. McKinsey calls agentic AI "as profound a structural shift as the advent of online commerce." The question for luxury brands is not whether to engage with this shift — it is whether they will still control their customer relationship when it arrives.

Imagine a scenario that is already happening. A customer thinking about a gift opens ChatGPT. They type: "I want to buy something special for my partner — she loves minimalist jewellery, classic leather, has good taste. Budget around €2,000." The AI responds with specific recommendations, explains the materials and brand story, and offers to complete the purchase. The customer never visits a brand's website. The customer never enters a boutique. The customer never sees the campaign that cost €3 million to produce.

This is agentic commerce — and it is not a hypothetical. ChatGPT accounted for 16% of Zara's inbound web traffic and 8% of H&M's between June and August 2025. Shopping-related AI searches grew 4,700% between 2024 and 2025. The front door of fashion retail is being rebuilt, and most luxury brands are watching it happen.

01 Why this shift is structurally different.

Every major disruption in retail has altered the discovery journey — the sequence of touchpoints between a consumer's first awareness of a product and their decision to buy. The internet moved discovery from physical browsing to search. Mobile moved it to wherever the consumer happened to be. Social media moved it to the content stream. Each transition required brands to rethink how they presented themselves at the new point of discovery.

Agentic AI is different in a way that matters especially for luxury. Previous disruptions added new touchpoints to the discovery journey — they did not eliminate existing ones. Agentic AI has the potential to replace the discovery journey entirely. When a consumer's question is answered, their options curated, and their purchase completed within a single conversation with an AI agent, the brand's website, the editorial photography, the carefully crafted narrative content — none of it is encountered. The brand has been intermediated out of its own customer relationship.

McKinsey / BoF State of Fashion 2026

~80%

Luxury consumers using AI tools for product discovery

McKinsey consumer surveys, 2025–2026

McKinsey Agentic Commerce, Oct 2025

McKinsey Senior Partner Holger Harreis, speaking in June 2026, framed the stakes directly: "Will the customer relationship and the transactions all move over to agentic conversational agents? If so, what will that mean for me if I'm a brand, a retailer, or a platform?" The answer to that question, he argued, is the most consequential strategic decision in fashion right now.

02 The luxury paradox.

The disruption is especially acute for luxury brands for a reason that is not immediately obvious: the very qualities that make luxury experiences valuable — personalisation, discretion, narrative, the sense of being known and served — are exactly what well-designed AI agents can simulate at scale.

A general-purpose AI agent that knows a customer's purchase history, understands their preferences, can explain the provenance of a Vacheron Constantin movement or the specific leather tannery behind a Bottega Veneta bag, and can do all of this at 2am without requiring a boutique visit — that agent provides something that most luxury brands' digital channels currently cannot. If the brand does not build this experience itself, someone else will build it for them — and the customer relationship will migrate to whoever does.

Luxury is not a business in which demand simply arrives fully formed. It is a business where meaning is made. The 'high-touch' layer — service, discretion, pacing, reassurance — is not a wrapping around the transaction. It is an essential part of it.

This is the luxury paradox of agentic AI: the shift is both the greatest threat and the greatest opportunity, depending on whether the brand is the agent or the product being recommended by someone else's agent.

03 Two scenarios — and what separates them.

Scenario A — Risk

The brand is cut out

The customer uses a multipurpose AI agent (ChatGPT, Gemini, Perplexity) to discover and purchase luxury goods. The brand's narrative, campaigns, and boutique experience are bypassed entirely. The agent recommends based on its training data — which may not reflect the brand's current creative direction, pricing rationale, or heritage story. The customer relationship migrates to the platform.

Scenario B — Opportunity

The brand owns the agent

The brand builds a proprietary conversational agent that knows the customer's wardrobe, syncs with their calendar, understands their taste, and proactively suggests — with the brand's voice, the brand's values, and the brand's narrative embedded in every interaction. The agent deepens the customer relationship rather than replacing it. Think: a personal stylist who is always available and always in character.

The brands already moving toward Scenario B are instructive. Kering launched KNXT with an AI assistant called Madeline — offering tailored luxury shopping experiences across its brand portfolio. The positioning is deliberate: not a chatbot for customer service, but a brand-authored curation experience. Daydream, a fashion AI startup partnering with 200 luxury retailers, has built an agent architecture where users' style preferences shape discovery across 8,000 brands — with the fashion brand's identity embedded in how recommendations are made and explained.

04 What this requires from luxury brands.

The strategic response to agentic AI is not primarily a technology decision. It is, as McKinsey has consistently argued, a business transformation decision. Most organisations that have invested in AI remain in the pilot phase — 88% of organisations now use AI in at least one business function, but only about one-third are scaling programmes across the enterprise. The gap between experimentation and transformation is where most brands currently sit.

For luxury brands specifically, the transformation required has three dimensions that are distinct from the general enterprise AI challenge:

1. 01

   Brand voice in the agent

   A luxury brand's AI agent cannot sound like every other AI agent. The voice, the vocabulary, the pacing, the level of detail — all of it must be brand-authored. An agent that responds to "tell me about this bag" with generic product specifications has failed the brand. An agent that responds with the story of the leather, the craftspeople, the creative direction — that is brand communication at scale. 02 First-party data as the foundation The customer relationship that an AI agent can deepen is only as good as the data that informs it. Brands that have invested in CRM, purchase history, preference data, and direct customer relationships will build better agents than those that have relied on third-party platforms. In agentic AI, first-party data is a competitive moat. 03 Presence in third-party agents Even brands that build excellent proprietary agents cannot ignore the general-purpose AI platforms where most discovery now begins. Being accurately and compellingly represented in ChatGPT, Gemini, and Perplexity — through content strategy, structured data, and brand narrative architecture — is the new SEO. Brands that do not manage this presence will be described, accurately or not, by someone else's training data. 05. The luxury-specific question There is one dimension of agentic AI that luxury brands must resolve that mass-market brands do not: the relationship between AI mediation and perceived exclusivity. Part of what makes a luxury boutique experience valuable is that it is rare, personal, and cannot be automated. If a brand's AI agent is too good — too available, too consistent, too frictionless — it risks undermining the scarcity and effort that luxury depends on. The resolution, for brands that find it, will likely involve using AI to extend and personalise the existing relationship with known customers, rather than to democratise access to the brand experience for new ones. The agent as a private stylist for existing clients, not as a public-facing discovery tool. This is the direction that the most thoughtful luxury implementations are moving — and it preserves the essential luxury logic while capturing the efficiency and personalisation benefits that agentic AI enables. Agentic AI will not destroy luxury. But it will redistribute the customer relationship — toward brands that understand it early enough to shape how it works, and away from those that treat it as a technology problem rather than a brand strategy problem. The window for being in the first group is narrowing.