Kitchen & Bath Business April 2021 - 22

TECH SAVVY
AI is the heart of the Plum Wine system, and
Hansen describes Plum as more of a software company than a hardware company. By matching the
label image to the winemaker,an incredible amount
of information is provided to the software and user
via the 7-in. touchscreen. Each time a user interacts
with the app, that information is captured so ML
learns user preferences on an ongoing basis.Another
feature is Plum's ability to share and compare favorites with family and friends via the web app. It can
make recommendations based on user feedback by
indexing the characteristics in each bottle of wine.
At one time,sommeliers were able to get their heads
around the characteristics of most wines that were
produced,but today that's not possible.This very immense and complex wine inventory opened the door
for Plum to design and build its virtual sommelier.

Appliances That Learn

The perfect bottle of wine selected with artificial intelligence

THE NEXT FRONTIER FOR MACHINERY MANUFACTURING is to design and
build machines so they can learn. Machine learning (ML) is the ability to
learn, in an ongoing way, how to perform a task instead of being programmed
to do a task in a specific way.
Some of the most innovative uses of ML are coming from the wine industry in production, distribution and preserving appliances, so I spoke with
Andreas Hansen, global president at Plum Inc., the manufacturer of an
automatic wine-dispensing appliance powered by artifical intelligence (AI);
Jack Palazzolo, head of marketing for Signature Kitchen Suite (SKS), the
luxury appliance division of LG and manufacturer of smart wine columns;
and Pam Dillon, CEO/co-founder of Wine Ring, an AI-driven technology
that helps consumers discover their wine preferences.

ABOVE LEFT Plum automated
wine-dispensing system
ABOVE RIGHT SKS wine preser-

vation unit

BELOW Wine ring image capture

MACHINE LEARNING BREAKS THROUGH

Complex things are the best opportunities for AI, and wine is one of the most
complex consumer products in the world. Wine has more than one million
types, 400+ individual characteristics and is a $300-billion global industry.
Dillon believes that the wine industry is a perfect entry point for ML because
it's the most complex sensory consumer product, and it's the second-most
complicated liquid on the planet. In case you are wondering, blood is first.
Her thinking is that if you can make a breakthrough in the application of AI
in the wine industry, then it's not going to be that difficult to do tea, coffee,
cheese, spirits and more.
DELIVERING WINE BY THE GLASS

If you like to enjoy a bottle of wine over a week or two without it spoiling, you
might be interested in Plum, an automated-wine dispensing system. According to Hansen, you simply open the door of the appliance, drop in a bottle or
two, close the door and then the magic starts. Using computer vision, Plum
takes a picture of the label and matches it to the winemaker.The wine is then
chilled to the temperature specified by the winemaker, and a spinning needle
pierces the cork or cap and fills the bottle with argon gas. No oxygen has
entered the bottle so there's no oxidation (the chemical reaction where wine
loses its vibrancy and eventually turns into vinegar). Each bottle is perfectly
preserved for up to 90 days, so whenever it's time for a glass, just choose your
pour size and with one touch, voila.

22
22_KBB_0421_TechSavvy.indd 22

SCOTT KOEHLER is
a 27-year veteran of
the kitchen and bath
industry and owner of
Dream Kitchen Builders, a design-build firm
doing business in North
Carolina.

INSPIRED BY WINE CAVES

Some of the earliest wine caves in the U.S. were
built in Sonoma and Napa, Calif., in the late
1800s. Wine caves are the ultimate way to preserve wine since they have the perfect temperature,
humidity and light (or lack thereof ). According
to Palazzolo, when SKS set out to make a perfect
wine preservation unit that controls light, vibration, temperature and humidity, they wanted to
replicate the way wine would have been aged in
wine caves in France.
SKS also collaborated with Wine Ring to add a
layer of AI to their system. Dillon notes that Wine
Ring works in the same context of a consumer
tasting wine and telling the sommelier what they
like and don't like.The software learns with interactions just like humans do - within the context
of the dialogue. Dillon calls her software dynamic
programming because not only does it capture
the info from the user, but it learns from it and
continues to do so.
Palazzolo says the biggest benefits of the ML
software are that it allows wine to taste as brilliant
as it can taste, and it helps make sure no bottle
of wine goes to waste. The SKS interface snaps a
photo of the label when the bottle is loaded. The
wine's characteristics are catalogued in the software, which allows users to organize their wine
and view the current inventory.The system learns
a user's preferences, makes recommendations and
can connect socially with family and friends via
the app to see what they like.
We're witnessing the early days of a whole new
era, an era in which machines can learn. It's great
to see appliance manufacturers break through
and add a layer of intelligence to improve our
lives. Machine makers of this new era will likely
be thinking of themselves more as software companies than hardware companies, just as Hansen
thinks of Plum.


/ APRIL 2021 / KBBONLINE.COM / The official publication of NKBA (NKBA.org) and KBIS (KBIS.com)

3/22/21 5:36 PM

http://www.KBBONLINE.COM http://www.NKBA.org http://www.KBIS.com

Kitchen & Bath Business April 2021

Table of Contents for the Digital Edition of Kitchen & Bath Business April 2021