Home Wine Business Editorial Technology Predicting Palates: Can Artificial Intelligence Improve Wine Buying?

Predicting Palates: Can Artificial Intelligence Improve Wine Buying?

1738
0
Advertisement

New technology is helping consumers make smarter wine choices.

By Gordon Feller

If you’ve ever felt intimidated when choosing a bottle of wine, you’re not alone. Today, a growing group of companies aim to demystify the wine experience through artificial intelligence (AI), tools that analyze the chemical compounds in any given bottle of wine. The goals are simple: to build a flavor profile and to offer personalized recommendations.

Tastry founder/CEO Katerina Axelsson
Tastry founder/CEO Katerina Axelsson

AI-powered Tastry helps wineries and retailers understand which consumers are going to like which wines — on an individual, local and regional level.  The company uses proprietary algorithms, collectively referred to as TastryAi, to understand what wine drinkers will like and then provide accurate recommendations. One algorithm interprets Tastry’s novel chemistry data, which pulls more than 5 million data points out of a single bottle of wine. “Understanding how that is going to taste, smell and feel to a human palate is the first step,” says company founder/CEO Katerina Axelsson. “It essentially reduces all the chemistry of the wine to a flavor represented mathematically.”

Another algorithm looks at that flavor, and the flavor of every other wine or subset of wines (such as 2,000 wines in a grocery store inventory), “and then chooses a set of simple questions that function as analogs to many compounds, or groups of compounds, found in wine,” Axelsson continues. 

(Image courtesy Tastry)
(Image courtesy Tastry)

One question may be, “How do you feel about the smell of fresh-cut grass?” The algorithm already knows which compounds, or groups of compounds, are responsible for that sensation, and narrows its recommendations accordingly. Says Axelsson, “envision AI as building a Venn diagram of questions that granularize the AI’s understanding.”

At this point another Ai-focused algorithm collates the information, essentially measuring the distances between the many vectors of the consumer’s palate and the available wines. Wines that are “closer” to the palate, are more likely to be good matches, and those that are further away are less likely to be enjoyed. 

Some retailers provide the TastryAi rating system, which lets consumers rate the products they’ve tried. Tastry has also built AI that pairs wines with foods, provides the best matches for groups of people, and addresses other common scenarios. Again, consumers answer a simple quiz and an entire retail assortment is ranked and matched. 

Retailers use Tastry, first and foremost, to provide informed recommendations to shoppers, thereby providing a better consumer experience. Retailers also use Tastry to identify new products. Axelsson says that, “once TastryAi provides recommendations to a substantial portion of a consumer base [this normally takes about 90 days], TastryAi can compare the retailer’s consumer palates on a store, local and regional level to better curate product mix in the store.”  

Improve winemaking and consumer choice

Alexandre Remy, Atlas Wine’s managing partner ​and winemaker
Alexandre Remy, Atlas Wine’s managing partner ​and winemaker

Since the wine industry is so steeped in tradition, says Alexandre Remy, Atlas Wine’s managing partner ​and winemaker, “it’s typically been averse to adopting technology for fear it eliminates the romance in winemaking.  

“I am all for technology,” he continues. “I think we need to be open-minded to all new technology and try our best to understand how it can help us to make better wines.” 

Winemakers are always challenged to produce the best wines every year; usually, it’s all about consistency. In that scenario, Remy thinks, “Tastry can use its computing power and AI technology to reduce the number of errors between vintages by providing year-to-year comparisons. Having more information for the winemaker is key to making the safest decisions and bringing to market the best wines possible.”

Another power of AI is to “trim” the work. When winemakers are blending, Remy says, “the number of samples and the diversity of volume can be overwhelming. AI can help pre-triage the samples to give more focus to the winemaker on the final touches.”

Remy also sees a role for AI when interacting with consumers. “Sommeliers are the bridge between wine and a consumer,” he says. “With their knowledge, a somm can guide the consumer to match the best wine for their occasion. AI could provide more information about consumers and also extend the spectrum of wine choices. Data from a company like Tastry adds a layer of expertise and knowledge to the sommelier, which can result in better matches for the consumer.”

Many voices making choices

Changing the way millions experience wine is not an easy undertaking. One approach is “review aggregation” — that is, sharing reviews, opinions and knowledge across a broad spectrum of wine lovers. 

Heini Zachariassen created Vivino as a solution after “feeling lost” whenever he’d face the wall of wine in a supermarket. Recognizing he wasn’t alone in this anxiety, he developed a solution that the casual wine drinker could use to pick a great wine at any price point. Vivino’s original vision was for consumers to know exactly what a wine tasted like before they spent a penny. 

Vivino CEO Olivier Grémillon
Vivino CEO Olivier Grémillon

Vivino’s strategy is to leverage what it calls “honest ratings and tell-it-like-it-is reviews” from its community of 60 million (the largest wine community in the world). According to Vivino CEO Olivier Grémillon, the company’s been “building a community for more than a decade to help combat [the intimidation factor].” He further says the “secret sauce lies in community wisdom coupled with machine learning.” 

Vivino harnesses the power of crowd-sourced information and technology. This strength in numbers, coupled with collected user data, makes wine easy to understand. Its two most prominent personalization features are “Match for You” and “Taste Characteristics.”

(Image courtesy Vivino)
(Image courtesy Vivino)

Match for You predicts the likelihood that a user will enjoy any specific wine. Says Grémillon, “The more wines a user rates and reviews on our platform, the more our algorithm learns about their unique tastes and can more accurately predict which wines they will and will not love.”

The Taste Characteristics feature aggregates the most commonly used words to describe wine, so that users will know just what they’re getting. As Vivino’s corporate marketers like to say: “No more fancy jargon getting in the way.”  

Making Room for Skepticism

Jason Cohen, CEO and founder of Gastrograph Corp.
Jason Cohen, CEO and founder of Gastrograph Corp.

Jason Cohen, CEO and founder of Gastrograph Corp. is “generally skeptical” of claims involving predicting consumer perception based solely on chemical data, because “individuals have large variations in their sensitivities, past tasting experiences, and adaptations from consumption choices. Any chemical information unfiltered through the cognitive filter of human perception is unable to predict sensory-preferences for a population or individual.” 

He considers review aggregators such as Vivino to be the first generation of consumer choice technology. “Individuals who know they prefer, for example, Chablis, can look at Chablis ratings of wines at different price points and at best find something they like, and at worst avoid obviously bad wines,” he says. 

Next-generation technology includes the ideal of personal recommendations. Cohen’s view is that, “the majority of ‘personal’ recommendations are simply consumer cohorting.” He finds some of these technologies interesting but, he believes, direct predictions of untasted products using chemical data is “unlikely” to provide any benefit over existing recommendation technology.

The next wave of technology, Cohen says, “encompasses both the cognitive filter of human perception and the driver of preferences of individuals.” His own company’s technology, Gastrograph AI, “models human sensory perception of flavor, aroma and texture to predict consumer preference of food and beverage products.” Gastrograph’s proprietary algorithms use AI and what techies call “machine-learning” to provide predictive analytics for the food and beverage industries.

The next wave of wine industry tech-innovation will be fascinating to watch. Some of the hype will not deliver, but it’s already clear that some significant software innovations could very well provide merchants, winemakers and consumers with better choices.

____________________________________________________

Gordon Feller

Gordon is a Global Fellow at The Smithsonian Institution’s Wilson Center. To research his nearly 500 published magazine articles he’s traveled to more than 55 countries. During his years at Columbia University (NY) he served as a Wallach Fellow, a Dean’s Fellow, and a Governor Lehman Fellow. Gordon received the Abe Journalism Fellowship, a special award created by the family of Japan’s Prime Minister. He formerly edited “Urban Age Magazine,” a periodical founded by The World Bank.

Advertisement

LEAVE A REPLY

Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.