2026-09-08
Young adults are increasingly turning to artificial intelligence for help choosing wine in restaurants, supermarkets and wine shops, according to reports in France and Spain that suggest digital tools are starting to take over a role long associated with sommeliers and retail staff.
Le Figaro reported this week that about 30% of French consumers have used AI when buying drinks, with the habit especially common among younger people. The newspaper, citing a French study published earlier this year, said 58% of people ages 18 to 25 have used AI to get information about an alcoholic beverage. Among those ages 26 to 35, the figure was 41%.
Restaurant owners and sommeliers told the newspaper that many customers now reach for their phones instead of asking for advice at the table.
“Young people prefer to ask ChatGPT,” restaurant owner and former sommelier Matthias Alberghi told Le Figaro. He said the appeal is simple: customers receive a quick and direct answer. Alberghi said the practice is now especially common among diners in their 20s.
The change is not limited to restaurants. Consumers are also using AI in supermarkets and wine stores, where they can upload photos of bottles or ask for food pairings on the spot. A 32-year-old shopper identified by Le Figaro as Valentin said he used ChatGPT while shopping for a wine to serve with raclette. He told the newspaper that he photographed several bottles and asked the chatbot for advice, adding that it even identified a grape variety it said would work best with the dish.
Similar reports have emerged in Spain. La Vanguardia said sommeliers in Barcelona were seeing more customers consult AI platforms to decide which wine on a restaurant list to order, rather than discussing their preferences with a wine professional. The newspaper described a growing pattern in which diners rely on a phone-based answer instead of a conversation that might include questions about taste, budget, food pairing or the style of wine they usually enjoy.
Judit Ayago, a sommelier at the Barcelona restaurant Masta, told La Vanguardia that some diners are drawn to AI because they feel uneasy about wine and worry about making a mistake. She said the wine business shares some responsibility for that discomfort by presenting wine in terms that can seem overly technical or distant.
Ayago said it troubles her that some customers treat an AI-generated answer as final and more trustworthy than a human recommendation. She said the trend reflects a lack of trust in sommeliers, even though wine service is supposed to guide guests through personal preference rather than deliver a single correct answer.
Other professionals were more dismissive of the idea that AI could replace them. Nataly Rodríguez, sommelier and co-owner of 55 Pasos in A Coruña, told La Vanguardia that AI would not do the job better than a trained sommelier. But even among critics of AI, some acknowledged that the profession has weaknesses that may be pushing customers toward technology.
Jofre Farran, of the Barcelona restaurants Alkimia and Alkostat, told the newspaper that AI use in restaurants is likely to grow because some sommeliers do not communicate well with guests. He said some wine professionals can seem more interested in displaying expertise than in asking questions and building trust. He also said the occupation has lost status in many restaurants, with too few sommeliers employed and too little value placed on the role.
The reports point to a broader shift in how younger consumers make purchasing decisions. AI tools can offer immediate pairing suggestions, explain regions and grape varieties, and narrow a list of choices in seconds. For customers who find wine culture intimidating or formal, that speed and privacy may be part of the attraction. Asking a chatbot can feel less stressful than admitting uncertainty to a server or sommelier in a busy dining room.
At the same time, the restaurant professionals interviewed by Le Figaro and La Vanguardia said the technology has clear limits. Alberghi told Le Figaro that AI can sometimes provide good recommendations. He said ChatGPT’s suggestion of a Crozes-Hermitage to pair with steak was strong enough that he might have made the same choice himself. But he also said other suggestions appeared to rely on broad assumptions rather than details about a specific bottle or vintage.
In one example, Alberghi questioned a recommendation of Chablis with grilled sea bass, saying the advice seemed based on a general image of Chablis rather than the actual wine available. He said that in wine retail settings he had seen the same problem more than once: the system often draws from professional tasting notes, but it can also produce information that is inaccurate or invented.
Le Figaro said it tested ChatGPT in a wine shop in Brittany and found examples that raised similar concerns. According to the newspaper, the chatbot first suggested a Quincy to pair with mussels, a recommendation that a former sommelier at the shop considered too expensive for the dish. It later suggested a 2025 Loire Chenin Blanc, but the wine professional said the system appeared to be extrapolating because it lacked detailed tasting notes for the latest vintage.
The newspaper said that when it checked the sources behind the recommendation, the problem became clearer: without enough information on the current vintage, the AI filled in the gaps. Le Figaro described that tendency to improvise when information is missing as one of the technology’s main weaknesses.
That concern matters in wine more than in many other consumer categories because recommendations often depend on details that are hard to reduce to rules. Vintage variation, producer style, serving temperature, food preparation and a drinker’s own history with certain flavors can all shape whether a bottle feels like the right choice. A wine that fits one person’s taste or budget may not fit another’s, even if both are eating the same dish.
Alberghi told Le Figaro that the bigger issue is not only whether AI gets the pairing right, but what is lost when people let software do all the choosing. He said wine is connected to memory, smell and personal experience in ways that a computer can only describe from secondhand sources.
His concern reflects a tension now appearing in many parts of hospitality: customers want convenience, but service professionals say convenience can flatten experiences that depend on conversation and trust. In the case of wine, AI may reduce the fear of asking a question, but it may also reduce the back-and-forth that helps people discover what they actually enjoy.
For now, the reports from France and Spain suggest that AI is becoming a regular part of the buying process rather than a novelty. Younger consumers appear to be leading that shift, using chatbots both before they order in restaurants and while standing in front of store shelves. Sommeliers and retailers say the tools can be useful, but they also say the technology is changing the relationship between customers and wine professionals at the point where many buying decisions are made.