If like me you have a penchant for wine, then your palette will soon be guided by an A.I. Sommelier. Artificial Intelligence has crossed the threshold for recommendations when it comes to our tastes. Prior to recognising what delights the tongue, search engines like YouTube, Amazon and Spotify have led the way in personalising our likes. They rely on our sight and sound senses when we select a video, book or song. Our selections build digital fingerprints about what we find appealing. Then predictions are made on the basis of them.
However, wine and all other food and drink are primarily based on our taste buds. Each of us is unique and our physical tastes are very subjective. For instance, do you prefer wines that are; crisp, earthy, dense, oaked, silky, complex, or velvety? These are just a few descriptive examples. Yet, until recently, it would have been extremely difficult to map and distil these nuances into qualities. And then use them to repeatedly make recommendations. This was due to the need for an extensive volume of controlled testing and numbers of people with varied preferences. In addition to taking account of all the chemical components that form part of the wine creating process.
Nonetheless, Katerina Axelsson, a chemist who founded, Tastry, has solved the problem. She pierced through the maze of taste bud roulette, by teaching a computer to taste wine. Being immersed in the wine industry, she recognised that it lacked coherent decision making data. Her epiphany occurred when a wine critic reviewed the same wine that had two different labels. And then scored them differently. Katerina said,
“I had a hypothesis that you could objectify sensory characteristics by creating a new flavour and analytical chemistry methodology that would measure products the same way a human palate does. And that this objective data could provide predictive visibility throughout the supply chain.”
Voila! By harnessing A.I, analytics and taste chemistry methodology, Katerina was able to gain insights into the flavour composition of wines. In turn, this helps us as consumers to buy wines based on our favourite flavours. Tastry’s algorithms has a 93% accuracy rate in predicting how consumers will score a wine.
This taste bud technology is gaining traction in America. It is already outperforming other tried and tested wine buying methods, by 45%. Tastry, is currently being trialled in the U.K and will be rolled out across Europe, later this year.
For those of you in the wine hospitality sector, you’ll be able to download an app called Bluebird, to receive wine recommendations.