Page 110 - Fortune-November 01, 2018
P. 110
ar tifi c i al intelli gence + life and lei sure
CHANGING WILL THIS AD
HOW YOU SHOP MAKE YOU SMILE?
BRICK-AND-MORTAR MARKETERS DON’T ALWAYS
stores have a new calling: hit the mark—we’re looking
They’re perfect A.I. data col- at you and that Kendall
lection labs. Home Depot is Jenner spot, Pepsi—but
using data from millions of increasingly, they’re leaning
transactions to figure out on artificial intelligence to
what else you—the DIY-er make those misfires less
grappling with, say, a big likely. Emotion A.I. firm
kitchen renovation—might Affectiva says a quarter
need and provide detailed of Fortune 500 companies
home project guides as well use its tech in their creative
as hyper-targeted cross- development processes,
selling. Sephora has used testing the reaction to po-
A.I.-powered facial recogni- tential ads in A.I.-enhanced
tion by ModiFace (recently survey research. Affectiva’s
bought by L’Oréal) to help system, which has been
shoppers select that exact trained on images of 7 mil-
right makeup shade: The lion faces (and 3.8 billion
software analyzes millions facial frames) from 87
of other past users to better countries, decodes the
predict what will look good facial expressions of indi-
MONEYBALL 2.0 on you. And MIT-spinoff viduals—the tech identifies
20 specific ones as well as
Celect uses machine learn-
ing to forecast how shop- eight emotions, including
pers behave, determine “disgust”—moment by mo-
WHEN NHL SCOUTS looked at Sean Durzi (above) in what kinds of promotions ment as they watch ads.
2017, they decided to pass on the 19-year-old de- work better in what part of Media research giant
fenseman. Just one year later, Durzi went second the store, and figure out KantarMillward Brown,
round to the Toronto Maple Leafs. The difference where products should be which has deployed Af-
placed for optimal results.
fectiva’s product since
was powerful new A.I. software by Montreal-based As for that price check 2011 (30,000 ads’ worth)
startup Sportlogiqthat parsed terabytes of data in aisle 10? Walmart, for found Nike’s lauded Colin
to uncover his powerful playmaking ability—call one, has tested robots at Kaepernick ad scored smiles
it Moneyball 2.0. Sportlogiq is just one of several 50 stores that scan shelves at key points. “We were really
companies using A.I. to help teams spot the next for out-of-stock items, able to pinpoint the fact that
star. “It’s all about identifying talent in hidden products placed back in the it was Kaepernick’s message
pockets and finding that diamond in the rough,” wrong spot by custom- about sacrifice and dreams
says Cam Potter, cofounder of BrooklynDynamics, ers, and incorrect prices that triggered the positive
an Australian data-analytics company that has —all time-consuming response,” says Graham
and cumbersome jobs for
Page, the firm’s managing
worked with several Major League Baseball teams humans. Though retail- director. They also found
and even developed a machine-learning A.I. ers are tight-lipped about that viewers responded
system for the 2017 Tour de France that collected in-store tech, companies positively, and with surprise,
real-time data points and spit out race predictions. like Navii and Simbe that to women players featured in
Brooklyn Dynamics is developing an app make A.I.-powered robots World Cup ads.
that will soon allow time-crunched scouts and are attracting attention and Page noted that beyond
coaches to run machine-learning analytics on investors, according to CB helping clients sharpen
both prospects and current players, creating a Insights. —Phil Wahba their campaigns, Kantar has
centralized database that can be accessed by gained broad insights that
college and pro teams around the world. “It’s a benefit all clients. Ads that DURZI: DAV ID K IROUAC —ICON SPOR T SW IRE V IA GE T T Y IM AGES
viewers describe as “pro-
unique tool to add to the recruiter’s repertoire,” gressive,” with protagonists
Potter says. “Other members of the organization featured in modern roles
can then look at [the stats] and join in on the draft (rather than traditional ones)
discussion to eventually decide who’s going to are 25% more effective, ac-
bring value to the club.” —C.K. cording to the firm. —E.F.
106
FO R T U N E. CO M // N O V. 1 . 1 8