FORBES ASIA
PING AN INSURANCEGENERAL GROWTH PROPERTIES US 1050T
GENERAL INS CORP OF INDIA IN 1624O
GENERAL MILLS US 499T
GENERAL MOTORS US 439T
GENERALI IT 136S
GENTING MA 1811 T
GENUINE PARTS US 928T
GENWORTH FINANCIAL US 904S
GEORGE WESTON CA 718T
GF SECURITIES CN 787T
GFNORTE MX 548 S
GILEAD SCIENCES US 197T
GIVAUDAN SZ 1199 S
GJENSIDIGE FORSIKRING NO 1653T
GKN UK 1064 S
GLAXOSMITHKLINE UK 226 T
GLENCORE INTERNATIONAL SZ 64S
GLOBAL PAYMENTS US 1387S
GOLDCORP CA 1274 S
GOLDMAN SACHS US 60T
GOODMAN AU 1587T
GOODYEAR US 1434 T
GPT (DIVERSIFIED FINANCIAL) AU 1910T
GRAND INDUSTRIAL CN 1828O
GRASIM INDUSTRIES IN 1076S
GREAT EAGLE HK 1546O
GREAT WALL MOTOR CN 803T
GREE ELECTRIC APPLIANCES CN 294S
GREENLAND HOLDINGS CN 341T
GREENTOWN CHINA CN 1547S
GRIFOLS SP 1122 S
GROUP 1 AUTOMOTIVE US 1976T
GRUPA PZU PL 770S
GRUPO ACS SP 532T
GRUPO AVAL CO 774T
GRUPO BIMBO MX 1345T
GRUPO BOLIVAR CO 1720T
GRUPO DE INVERSIONES
SURAMERICANA CO 1674 T
GRUPO GALICIA AR 1785S
GRUPO INBURSA MX 1073T
GRUPO MEXICO MX 581S
GRUPO TELEVISA MX 1695T
GRUPO ZULIANO VZ 1914T
GS ENGINEERING KO 1819S
GS HOLDINGS (CONGLOMERATE) KO 1032S
GUANGDONG INVESTMENT HK 1597S
GUANGD WENS FOODSTUFFS CN 973T
GUANGZHOU AUTO CN 691S
GUANGZHOU R&F CN 725S
GUANGZHOU RURAL
COMMERCIAL BANK CN 1092O
GUDANG GARAM ID 1494T
GUNMA BANK JA 1620T
GUOCO HK 1608S
GUOSEN SECURITIES CN 1042T
GUOTAI JUNAN SECURITIES CN 696T
H LUNDBECK DE 1890O
H&M SW 583 T
HACHIJUNI BANK JA 1590S
HAINAN AIRLINES CN 1108S
HAITONG SECURITIES CN 730T
HAKUHODO DY (MEDIA) JA 1961T
HAL TRUST MC 1024T
HALKBANK TU 873 S
HALLIBURTON US 743 T
HALYK BANK KZ 1595O
HANA FINANCIAL KO 436S
HANG LUNG HK 1501T
HANKYU HANSHIN JA 1115T
HANWA JA 1760 S
HANWHA KO 851 T
HANWHA CHEMICAL KO 1420T
HARBIN BANK CN 1154T
HAREL INS INV & FIN SERVICES IS 1821O
HARRIS US 1191 T
HARTFORD FIN SERVICES US 656T
HASEKO JA 1647 S
HCA US 275T
HCL TECHNOLOGIES IN 936S
HCP US 1988 T
HDFC IN 321 S
HDFC BANK IN 202S
HEIDELBERGCEMENT GE 496 T
HEILAN HOME CN 1938O
HEINEKEN NE 390T
HELLA KGAA HUECK GE 1803S
HELLENIC PETROLEUM GR 1736S
HELVETIA SZ 1095 T
HENDERSON LAND HK 607S
HENGAN INTL CN 1755T
HENGLI PETROCHEMICAL CN 1833T
HENKEL GE 293 T
HENRY SCHEIN US 1275T
HERMÈS INTERNATIONAL FR 844T
HERO MOTOCORP IN 1583S
HERSHEY US 995 T
HERTZ GLOBAL US 1519T
HESS US 1238 T
HESTEEL CN 1212 T
HEWLETT PACKARD ENTERPRISE US 316T
HEXAGON SW 1255S
HIKVISION CN 835 S
HILTON US 775 T
HINDALCO INDUSTRIES IN 1186T
HINDUSTAN PETROLEUM IN 951T
HIROSHIMA BANK JA 1601T
HITACHI JA 149 S
HNA TECHNOLOGY CN 1231T
HOKKAIDO ELECTRIC POWER JA 1917T
HOKKOKU BANK JA 1795S
HOKUHOKU FINANCIAL JA 1543T
GLOBAL 2OOO
SUP TDOWN
OUNCHANGED ONEW
ASIAN COMPANIES ARE IN RED TYPE
financial institution in the world has this: 436
million internet users, 166 million financial
services customers.
What’s behind Ping An’s transformation
into a technology-based insurance group?
How did it get started?
Five years ago we already had our own pro-
prietary facial recognition team; four years ago
we started working on voiceprint. Whatever
the language, Spanish or Sichuanese, as long as
we have 20 seconds of your voice, we can trans-
late it into a mathematical matrix as unique as
your thumbprint. Our brain can only remember
the voices of about 150 people. Our system can
recognize unique voices of 100,000 people.
In the last three years, our system has
been trained to recognize micro-expressions,
the tiniest movement of our eye muscles and
around the lips, within milliseconds. We use
this in our real-life loan interview done on the
mobile app by our credit assessors who ask
random questions just to see if a borrower is
lying—more than 80% positive prediction rate.
As a financial services
and healthcare provider,
we need to verify who
the customers are when
everything is done on the
mobile. The MIT com-
puter science-artificial
intelligence lab dedicated
a professor to work on a
project with us over the
past four years.
How do you pick businesses to focus on?
A big problem in China is the lack of aford-
able, good-quality primary care, very diferent
from Taiwan, Hong Kong or Singapore, where
one can just walk into a medical center to see
a doctor.
O ine, we built a technology platform that
set standards for outside clinics—65,000 of
them now—to standardize customer records,
set pharmacy and medication and procure-
ment standards, with links to social security.
In the city of Chongqing, where over
60% of people are likely to contract chronic
inflammatory lung disease, with the ministry
of health’s in-depth study of health records
we built an AI-powered system to predict the
likelihood of a person of getting this disease
based on records, including home-life condi-
tions, such as a house with an open chimney(three times more at risk) as opposed to a
modern ventilation system.
Much of the group’s technology is home-
grown. How have you managed to pool tal-
ent, the right strategy and execution? Do
you have an incubation center in Shanghai?
We no longer have an incubation center. We
have top-down and bottom-up approaches.
The top-down is at the beginning to identify
where we want to be. Then we set up a com-
pany. The bottom-up is the management team
tasked to get the business of the ground. We
close down underperforming companies.
We can bring a lot to auto insurance—we
are China’s second-largest auto insurer and its
largest auto-dealer financier and the largest in
car financing. We bought New York-listed Au-
tohome two years ago, the only business we did
not build from scratch. We turned it into a car
portal from an advertisement media company,
increasing its size by 3.5 times in two years.
In smart cities, we worked in Shenzhen to
ease traic congestion. Last year traic jamsduring the Golden Week lasted 21.7 hours.
This year it was zero, after the Shenzhen traf-
fic police used our technology for congestion
management.
You have a core competence in handling
city traic?
Our insurance claims ratio is the lowest
in the industry, helped by our AI capabilities
covering 360 Chinese cities. In a car accident
in China, drivers—a lot of them first-time
drivers—have to stay there for traic police to
handle the liability allocation.
Our investigators would arrive at the site
within five to ten minutes 90% of the time,
even faster than an ambulance or police—no
small feat and only possible with technology.
Our AI capabilities can predict where acci-
dents are likely. We analyze the past 20 years
of auto claims for what the likelihood of48 | FORBES ASIA JUNE 2018“OUR INVESTIGATORS WOULD
ARRIVE AT THE ACCIDENT WITHIN
FIVE TO TEN MINUTES 90% OF
THE TIME, EVEN FASTER THAN AN
AMBULANCE OR POLICE.”