全球奢侈品：人工智能悖论

Global Luxury Goods: The Artificial Intelligence Paradox Luxury goods companies are selling craftsmanship, creativity and exclusivity. An overt andenthusiastic adoption of artificial intelligence puts their core promise and credibility at risk.Nevertheless, we are seeing major new developments that investors should be aware of.The path to success is more in the ability to capture data and structure internal processes -where 80% of the value would be created - rather than in the algorithms, which would beavailable to all. This would once again favor scale and large organizations. We thank BCG fora constructive exchange of views on this topic Luca Solca+41 582 723 126luca.solca@bernsteinsg.com Maria Meita+44 20 7170 0540maria.meita@bernsteinsg.com Eric Chen, CFA+852 2123 2628eric.chen@bernsteinsg.com Back office problems have been the first step for artificial intelligence applicationsin fashion and luxury goods. The fashion and luxury goods industry faces highlycomplicated (but not conceptually difficult) supply chain optimization problems when itcomes to inventory planning, deployment, and replenishment (see Global Luxury Goods:Getting Smarter... the Artificial Intelligence Opportunity). Large brands have severalhundred directly operated stores, located across multiple customs regimes, and manythousand SKUs. Before artificial intelligence, these problems had been tackled with linearprogramming and machine learning algorithms = applications in this field have beenaround for a while with a skew to number crunching. Other back office problems firstin line to be attacked have related to trend detection and analysis: social media providemillion of “data points” that artificial intelligence applications can attack far faster andmore efficiently: 1) Demand forecasting, 2) Trend analysis from fashion images andonline searches, 3) Inventory management, 4) Logistics optimization, and 5) workforceoptimization. Yi-Peng Khoo, CFA+44 20 7676 6822yi-peng.khoo@bernsteinsg.com Specialist Sales Alix Turner+44 20 7762 4044alix.turner@bernsteinsg.com A second wave of artificial intelligence applications has developed to support therise of online sales - with a major boost through the Covid-19 pandemic(see GlobalLuxury Goods: Strategic implications of the "digital revolution"). The idea here has beento provide personalized navigation and suggestions based on browsing history, as wellas to allow consumers to find products on the back of searching for images rather thanwords. The effort has been to capture the largest number of data points from consumersinteracting with brands online, and - ideally - to integrate this data with other collectedfrom in-person interactions: 1) Recommendation engines, 2) Visual search, 3) Customersegmentation, and 4) Predictive marketing. A major threshold was broken by introducing artificial intelligence at the interfacewith the customer, serving as “digital sales assistant”.This has gone from stylingassistants suggesting how to mix and match different products, as well as chatbotsanswering questions, virtual try-on applications - both in store (through “intelligent mirrors”)and online. The introduction of AI at the customer interface marks a shift from assistedselling to AI-mediated journeys, where discovery, consideration, and transaction canincreasingly happen within generative interfaces. This raises strategic questions arounddiscoverability (GEO), brand control, and the role of agentic commerce, requiring luxurybrands to actively design their presence within AI ecosystems. The effort has been to makethe client interaction more relevant with the goal to increase conversion: 1) Chatbots andconcierge assistants, 2) Virtual try-on, and 3) Styling recommendation engines. Continued on the next page… … continued from the first page The next step (which is in progress) is to bring artificial intelligence at the core of the4Ps of client facing marketing. •Consumer CRM and dialogue.AI models can produce and sustain a relevant dialoguewith consumers, replacing / supporting human intervention and massively expandingthe number of people receiving personalized treatment. For both online and in-store experience, AI models can help design customer-centric layouts and visualmerchandising strategy that’s aligned with product performance and commercialpriorities (see Retail meets Darwin: The Traffic Challenge). •Marketing communication.AI models can generate high-grade quality contentfor social media and communication at scale, with natural sounding translation forlocalization. Internally, AI models can also help dynamically optimize marketing spendallocation across regions and channels based on real-time performance data (see GlobalLuxury Goods: Social Media mini-series - Strategic background). •Collection articulation and merchandising.AI models can start from core creativeinputs and “fill the gaps” in a collection, speeding and supporting key merchandisingprocesses (see Global Luxury Goods: The