top of page
  • Writer's pictureStartU HQ


Coqovins – Your Personal Virtual Sommelier

Coqovins is a Wharton startup utilizing AI to provide personalized wine recommendations for customers. The company relies on machine learning to capture customers’ wine preferences and aims to help wine retailers increase sales by facilitating customers’ purchase decisions. I sat down with the Coqovins Co-Founder, Quentin Chalvon Demersay, and discussed his take on combining AI and wine expertise to transform a traditional industry, thereby creating a win-win solution for both retailers and customers.

More options, more problems

Coqovins, the virtual wine sommelier, is helping customers choosing the wine that best fits their preferences.

Co-founder Quentin Chalvon Demersay, aims to infuse data analytics into the traditional wine industry. The Wharton MBA ’18 and former Facebook data scientist has long held an affection for the subtleties of wine – drawing inspiration from his family roots in Burgundy, France and the 200 year-old wine producer Louis Latour. Quentin, together with his co-founders Enguerrand Chalvon and Edouard Dole, identified a common issue that most wine consumers face when they enter the wine store – the overwhelming abundance of choices.

More options, more problems. Photo Credit:

People vary significantly in their tastes and preferences for wines, regardless of their price. The problem is, when the customers enter a brick and mortar wine store, they are presented with on average 200-300 different wine choices. However, there isn’t enough helpful information on the bottles that can be easily decoded by the customers. According to Quentin, the staff also isn’t helpful and over 50% of customers wind up using their mobile device to get better information. As a result, most people decide to buy the second cheapest wine and they also buy less – Coqovin’s market research suggested that 26% of the time customers leave the store empty-handed.

Coqovins – Your Personal Virtual Sommelier

Users start by answering questions such as the price they are willing to pay or the occasion for which they will be consuming the wine. These questions are developed by 2 human sommeliers that work in France to ensure the comprehensiveness of the question list. Based on their answers, AI will determine users’ wine profile and then combines data analytics and sommelier expertise to make the most relevant wine recommendations. After the wine is consumed, users are asked to give a 1 to 5 rating to the recommendations to help the algorithm make better choices in the future. Think of it as the “movies you might like” section on Netflix, where it uses the machine-learning algorithm to suggest movies according to customers’ past viewing habits.

Coqovins, Your Personal Sommelier

Although the algorithm sounds daunting, the user experience is quite simple – all a user needs to do is to install Facebook Messenger in order to access the app. When customers enter the store, they can scan the big Coqovins QR code displayed in the wine store shelf. The store-specific QR code is associated with the wine offerings in that store. A virtual sommelier chatbot will pop up in the Facebook Messenger app and ask users simple questions related to their preferences. The user will then receive recommendations based on the available wines in the store.

Wine-drinking Millennials

Tech-Savvy Wine Drinkers. Photo Credit:

Coqovins is targeting wine drinking millennials in the world’s largest wine market.

According to Statista, in the U.S., 132 million people drink wine and there are 19,651 outlets selling wine. The total wine retail sales are estimated to be $39 billion in 2016, which translates to $128 in annual spending per capita. Millennials represent 42% of the U.S. wine consumption and 73% of them are regularly using Facebook Messenger. Wine retail stores lose several billion dollars per year due to customer confusion. Coqovins’s virtual sommelier fills the gap and help retailers to recoup the losses.

Win(e)-win(e) Solution for Wine Retailers and Customers

But would a personal sommelier be too expensive? Not at all.

Coqovins Business Model. Photo Credit: Coqovins Pitch Deck

Coqovins uses a subscription model for retailers and doesn’t charge anything to the wine customers. Coqovins’ beta testing showed an increase of customer conversion rate and average spending by 20% to 40%. Better purchase decisions also lead to increases in customer loyalty. If the result proves to be consistent in a larger sample, Coqovins would be a win-win solution for both the store and the customers.

Other Virtual Sommeliers?

According to Quentin, currently most wine recommenders present a lot of information and use an average rating system, such as the one used by The former continues to inundate customers with information while the latter is inefficient in providing a recommendation that’s most suitable to individual unique tastes.  

Coqovins, on the other hand, provides actionable recommendations with wine available in the store. The artificial intelligence solution is scalable, channel agnostic, and adaptive.

Coqovins AI Power. Photo Credit:

To date, Coqovins is active in 5 wine stores in France and is exploring opportunities in the US. The company strives to become the leading virtual sommelier in the U.S. and to eventually present the app to restaurants, e-commerce and other business models.

bottom of page