This document discusses HDFC Life's use of vision AI technologies. It outlines several problems in the insurance customer journey like complex products and lengthy documentation. HDFC Life implemented several vision AI solutions like FaceAgeing, FaceSense, PicReader, Bodmeter and SelfieQuote to streamline processes like KYC, underwriting and claims. These solutions provided benefits like faster servicing, improved customer experience and resource savings for HDFC Life. Key results were 3000+ leads generated and 175+ new customers from SelfieQuote, with 92%+ facial matching accuracy from FaceSense reducing verification time. Overall the vision AI solutions helped improve customer engagement and provided revenue benefits for HDFC Life.
HDFC Life leverages computer vision to enhance customer experience and drive growth
1. This is the sole and exclusive property of HDFC Life. April 15, 2021
The use of Vision AI at HDFC Life
Presented by Francis Rodrigues,
Senior Vice President, Innovation, Data Labs and Business Insights , HDFC Life
2. 2
Established in 2000, HDFC Life* Insurance Company is a leading individual &
group insurance solution provider in India
3. The problem, the opportunity and the objective
3
Problems & Opportunities
Insurance is perceived as a
complex product by the
younger generation
The buying journey is
complex with lengthy forms
and documentation
Growth of computer vision
applications due to cheaper
computing capabilities
The rise of the selfie
generation -93 million selfies
clicked in a day according to
google
HDFC Life had 4 million
customer photographs
5. We identified key problems and opportunities in the customer value chain where AI
based solution could be explored
5
Pain
Points
Limited CSAT
KYC requirement
Dated photographs
on KYC
Manual query
handling
Manual Form filling
Error Prone
Lengthy form fill
Sign-in required
Human Judgment
Lack of medical
markers
Long waiting time
Inexperienced
attendants
Re - authentication
Dated photographs
on KYC
Missed payout
Recycling
Fraud
Underwriting Claims
Form Filling Servicing
Insurance Buying and Ownership Journey
Customer
Connect
Lengthy
Documentation
6. Various solutions using voice AI, vision AI, text AI and machine learning were
explored to solve these problems across the customer value chain
6
Opportunity
Text AI
Voice AI
Vision AI
ChatBots
Voice AI
Vision AI
ChatBots
Voice AI
Machine Learning
Machine Learning
Vision AI
Vision AI
ChatBots
Voice AI
Vision AI
Machine Learning
Underwriting Claims
Form Filling Servicing
Customer
Connect
Insurance Buying and Ownership Journey
Machine Learning
7. The use of Vision AI for the Policy Purchase Journey
7
AgeTymer
Face Ageing to visualize
the need for insurance
FaceSense
Face Verification with
customers pic at PCVC*
PicReader
Document OCR for
PAN, Aadhar &
Passport
Bodmeter
Helps to gauge
customers BMI from his
image
SelfieQuote
Insurance buying with
the help of a selfie
Attractive for younger
customers who wish to
buy quick insurance or
older customers who
want to buy online but
find the process
intimidating
Insurance being
intangible, this makes
the customer visualize
the risk of old age and
hence serves as a
nudge in the sales
journey
Extracts customer
gender, age & BMI and
lets us know if they are
overweight. This input
is used to flag off cases
where medical is
required
Extracts information
from standard ID proofs
submitted by the
customer thereby
eliminating the
requirement manual
extraction
Compares customer selfie
to the image on the
document & gives accuracy
of the match. Also matches
the customer credentials
from the ID documents to
do a KYC match at PCVC*
B
E
N
E
F
I
T
S
Policy Purchase Journey
* PCVC=Pre conversion verification check
8. The use of Vision AI for Policy Servicing, Risk Mitigation and Fraud Prevention
8
PicReader
Document Verification
at Claims & Payout
Signal processing based
Health Profiler which
shows the customer
Blood Pressure and
Heart Rate through a
15 sec video of his/her
fingertip by placing it
on the camera of the
smartphone
Face sense authenticate
the customer who has
walked into a branch
for a policy payout with
the photographs
submitted at the time
of policy inception
Compares the
submitted
document/scans with
the original documents
submitted to validate
the customer/claimant
details
B
E
N
E
F
I
T
S
Policy Servicing, Risk Mitigation and Fraud Prevention
HealthProfiler
BP and Heart rate
detection
FaceSense
Customer Verification at
Branch Payout
FaceMed
Customer Verification at
Medical Checkup
Face Med of the
Customer visiting the
medical center for a
check to authenticate
themselves with their
photos to prevent
misrepresentation
AI Life Certificate
Life Certificate Issuance
for annuitants
Annuitant records a 10
sec video which is
authenticated against
the policy details and a
life certificate is issued
to the annuitant
without them visiting
the branch
9. The use of Vision AI for Employee authentication during Covid19 WFH
9
Employee Authentication
With all employees forced to work from home due to the Covid19
pandemic, HDFC Life launched Work Smart, a face authentication
software which allows an employee to login to the company systems
using his face as a password.
This feature was launched to ensure secured access to the company
systems and applications by authorized personnel
Work Smart
11. These AI solutions have led to broad benefits for the customer and the business
11
BENEFITS
ACHIEVED
CUSTOMER BUSINESS
Real-time Authentication
Faster servicing
Futuristic perspective
Convenience
Smoother native
experience
Communication via
everyday apps
3000+ Leads & 175+
Customer
45.06 lac EPI
92% Face Match Accuracy
Manual Verification TAT
from 3min to real time
0.6 resources saved /
branch
20 call center seats
reduced
SelfieQuote - Initial pilot with our distribution partner generated 3000+leads translating into 175+ new customers and a business of INR
45.06 lakhs in a month with a reduced issuance TAT of 3 days
FaceSense has achieved a 92%+ accuracy in validating a person’s live image with his photo ID. This has saved the customer valuable time by
reducing the time required for physical id check from 3 minutes to near real time. The project has realised a revenue impact of 10 Cr.
HealthProfiler has been integrated in a customer facing retiral platform for calculation of a Health Score of the customer. The platform has
received a 23% MOM improvement in customer engagement rate
FaceSense at Branch
Live at 300+ Branches
Total Customers for facial
recognition:50,927
FaceSense at Medicals
Total Customers for facial
recognition:22,560
FaceSense at Chat PCVC*
Total Customers for facial
recognition:11,350
Life Certificate
Total Customers for facial
recognition:5000
85% success rate