This piece, authored by Anita Balchandani, Gemma D’Auria, Holger Harreis, Roger Roberts and Steven Begley, with contributions from Katharina Giebel, argues that AI is moving the decisive moment of shopping upstream—into the environments where intent is first expressed and framed. For decades luxury strategy centered on destinations: boutiques were the main front door and digital channels extended the brand’s curated world. Now, general-purpose assistants and shopping agents are increasingly the places where comparison, framing and early curation happen.
Why this matters for luxury
The change is more than technical: it affects growth and productivity in a sector where experience and service are part of the product. For ultra-high-value clients, AI promises adviser-grade continuity beyond the boutique. For the next tier of luxury consumers—many of whom will never have a named client adviser—AI can act as a scalable digital proxy, offering interpretation, curation and reassurance previously reserved for the top tier.
The strategic question is not whether AI will mediate luxury journeys but whether brands will govern that interpretation layer or let others define it. McKinsey’s broader research estimates agentic commerce could mediate $3 trillion to $5 trillion of global consumer goods commerce by 2030, underscoring the speed and scale of the challenge.
What McKinsey’s surveys show
The article draws on two McKinsey surveys: the Global McKinsey Agentic Shopping Experiences Consumer Survey 2025 (fielded December 16–22, 2025; n = 300; luxury segment n = 31) and the Global McKinsey Agentic Shopping Experiences Merchant Survey 2025 (fielded December 19–30, 2025; n = 90; luxury respondents n = 19). Given the small luxury subsamples, results are directional snapshots.
Key consumer findings for the luxury segment:
- 85% use multipurpose AI assistants (e.g., Google AI Mode or Perplexity) to support shopping decisions; 52% report frequent use.
- 74% have used visual search (uploading a photo to find a product or emulate a style), with roughly half using it frequently.
- 55% have tried virtual try-on tools; 15% use them frequently.
- 83% report “high” or “very high” satisfaction with AI shopping tools; only 4% are dissatisfied.
These figures show AI use among luxury shoppers is substantive and sustained. As a result, the first interpretation of desire increasingly occurs off-premises—in multipurpose assistants or emerging category agents—making the interface where intent is first interpreted a contested space.
Delegation, trust and limits
Luxury consumers’ comfort with agent involvement varies by stage:
- Discovery and selection: 50% comfortable.
- Transaction execution: 58% comfortable.
- Care (service, reassurance, relationship stewardship): 39% comfortable.
Preferences for overall delegation are also telling: only 9% prefer full autonomy; most opt for assist (32%), assemble (31%), or authorize within guardrails (28%). In other words, consumers are not rejecting agents but setting clear terms.
Top concerns among hesitant luxury consumers include privacy/data misuse (44%), overreliance on AI (40%) and accuracy doubts (34%); only 18% name distrust of AI companies as a top-three concern. When asked what would make them comfortable with agent-mediated interactions, luxury shoppers emphasize transparency about the agent’s role (54%), data-privacy safeguards (52%) and proven reliability (52%). These are prerequisites for letting agents operate near luxury-defining moments.
Consumers estimate that 39% of their luxury purchases will be mediated by AI agents by 2030; merchants expect an even higher share of interactions and transactions to be agent-mediated (47%).
How brands are responding
Across retail, experimentation with agentic interfaces is advancing. Luxury houses have so far been selective: rather than end-to-end agentic journeys, many are piloting interpretation and personalization layers to augment discovery, curation and clienteling while protecting brand meaning.
Examples cited in the research include:
- Ralph Lauren’s “Ask Ralph,” a conversational stylist inside the brand app.
- Brunello Cucinelli’s AI-driven environment that embeds house philosophy into conversations.
- Moncler and others testing AI-enhanced personalization and immersive engagement.
- Zegna’s ZEGNA X, a copilot and data ecosystem that augments sales associates’ clienteling productivity.
In McKinsey’s merchant survey (luxury respondents, n = 19), 68% point to "more personalized experiences" as the primary source of excitement; 39% cite smarter discovery; 32% omnichannel integration. The merchant posture is fragmented: 36% are bullish and wanting to lead (11 respondents), 29% cautious and monitoring (9 respondents), and 36% more skeptical (11 respondents). Only 4% (1 respondent) view agentic commerce as an existential threat; 82% (23 respondents) see it as an experience and efficiency opportunity.
Luxury merchants tend to prioritize interpretation over mere convenience: 68% flag personalization (versus 52% in specialty retail), while only 14% emphasize convenience/time savings (versus 29% in specialty retail). This selectivity reflects the high cost of getting interpretation wrong in luxury.
The strategic imperative: codify judgment
The authors argue the strategic prize in luxury is interpretive authority: defining the rules by which agents should understand, filter and present your brand. The instinct of a great client adviser—discretion about what to show, restraint about what not to show, reassurance at moments of uncertainty—must be translated into machine-readable policy and auditable behavior. Autonomy should reduce friction while preserving human judgment at moments where meaning is created.
A practical model uses four “rooms” (not linear stages) representing contexts where delegation creates or erodes value. The challenge is planning the transitions, ensuring reliable handoffs and preserving human backstops where commitment carries reputational weight.
Implications for stores and operating models
Upstream mediation raises the bar for physical retail rather than diminishing it. As agents handle more filtering and comparison, stores become loci of confirmation: fit, feel, proportion, authenticity and the human adviser’s reassurance. In-store AI is most powerful behind-the-scenes—surfacing context, anticipating needs and coordinating care—while leaving human judgment at the moment of commitment.
When done well, AI can raise both the floor and ceiling of clienteling: surfacing client context in the moment, converting upstream shortlists into brand-coded looks, checking eligibility and scarcity rules, and coordinating fulfillment and aftercare. But the cost of error concentrates at commitment, increasing the value of human backstops and service-recovery mechanisms.
The competitive threat is not that agents will replace brands, but that generic agents will erode brand meaning by reducing luxury to price, availability or popularity. If brands do not author how they should be interpreted, others will.
Conclusion
Agentic commerce will not change what luxury is but will change how decisions are made and how customers engage across occasions and categories. Consumers already begin journeys across a broad set of AI surfaces—search engines, personal agents and brand apps—so the front door is contested. Luxury winners will be those that make their codes legible to agents, build discretion and control into systems, and use technology to deepen rather than dilute relationships.
Authors: Anita Balchandani (senior partner, McKinsey London), Gemma D’Auria (senior partner, McKinsey Milan), Holger Harreis (senior partner, McKinsey Düsseldorf), Roger Roberts (partner, Bay Area), Katharina Giebel (consultant, Bay Area), Steven Begley (senior partner, McKinsey New York). Edited by Larry Kanter (senior editor, New York).
Note: The consumer survey was fielded December 16–22, 2025; the merchant survey December 19–30, 2025. The analysis complements McKinsey’s research on the agentic commerce opportunity and the automation curve in agentic commerce.
