Upload
igate-corporation
View
654
Download
2
Tags:
Embed Size (px)
Citation preview
March 13, 2015 Proprietary and Confidential - 0 -
Warranty Fraud Detection How connection, sensors and data can revolutionize detection & much more
Abhay Mahagaokar Manufacturing Business Unit Head & Senior Vice President
March 13, 2015
March 13, 2015 Proprietary and Confidential - 1 -
Happenings in our world
The Economy continues to be Robust
Focus on costs continue to be robust
Manufacturers seek to improve products & experiences
Newer business models are sprouting
$
Tech disruption underway
Sensor sales growing at 70% annually since 2010
Pervasive connectivity
High power, in-memory computing enabling Analytics
Tech
March 13, 2015 Proprietary and Confidential - 2 -
Warranty evolution..
Manual Claim process
Administratively Intense
Cost focused
Higher focus on Product Reliability
Warranty Management Platforms
Data Analysis & KPI’s
Process specialization
Reactive to Proactive…Hunch to data driven ’Cross the Chasm’
Reactive
Automated
Integrated
Improved Efficiency
Product reliability emphasized
Repair Database
Smart Products – embedded sensors
Connected infrastructure
Bring together multiple data Sources – claims, social, Machine
Predictive Analytics
‘Smart’ Warranty
March 13, 2015 Proprietary and Confidential - 3 -
Implications & opportunities
Warehouse & Distribution
Factory
Stores Market
Sense experience & Usage
Reduce Manufacturing Variances
Control Supply chain conditions
Talk to the Product
Data Security & Privacy
March 13, 2015 Proprietary and Confidential - 4 -
Taking the audit route to fraud
Case study – auto OEM
Manual… Qualitative… Time Consuming
March 13, 2015 Proprietary and Confidential - 5 -
Addressing fraud
Traditional Business Rules
Pattern identification
Valid Claim Warranty Claims
Claim Engine
Machine data
Fraud Analysis
Auto validated Claim
Pattern based learning
Past
Future
March 13, 2015 Proprietary and Confidential - 6 -
Addressing Fraud
Data extracted from devices, enterprise systems &
Real Time Platform for processing sensor data
Scalable & Flexible Platform
Historical Intelligence
Automate product registration and warranty period management
Device Provisioning
Streaming Analytics Platform
Clustering Algorithms to evaluate unstructured text
Text Analysis
Traditional Rule Mining for standard anomaly detection
Segmentation of historical event detection
Sequential Rule Mining for Pattern Detection & fraud detection
Machine Learning
March 13, 2015 Proprietary and Confidential - 7 -
The new warranty approach
Claims Customers Products Policy Campaign RMA Early Warning
Product Engineers
Analysts & Auditors
Service Engineers
Moving From Traditional Warranty Management Solution to an Approach Based on Real
“Product Data”
Infrastructure
Claim Contact Center
Processing Center
Settlements
End User | Retailer | Distributor
Reverse Logistics Provider
Client’s Back-end Intelligence Systems:
CRM, ERP, etc. Service Centers
March 13, 2015 Proprietary and Confidential - 8 -
The technology construct
Monitoring & Action
Visualization, Alerting
Sensor data analysis and Processing
Sensor Data storage and management Se
curi
ty
Sensor Data Acquisition
Sensors & hardware interfaces
Se
nso
r M
anag
em
en
t
Bil
lin
g
Pro
visi
on
ing
Business Architecture
March 13, 2015 Proprietary and Confidential - 9 -
The technology construct
Technical Architecture
IoT Gateway Layer IoT Middleware
MQTT-SN
TCP
Service Orchestration
MQ
TT
Me
ssag
e B
rok
er
Stream Analytics
Rule Engine
Data store Management
Event Filter
Middleware Gateway Agent
Data Aggregator
Dashboards
Security
Things/Device Management
Fraud
CAPA
Things DB Alarms/Rules
Machine Stream
Data warehouse
APIs
Apps
Enterprise system
Warranty system
Sensor
(Managed) End Node
Device
End Node 2
Co
nn
ect
or
Ag
en
t
Co
nn
ect
or
Ag
en
t
March 13, 2015 Proprietary and Confidential - 10 -
Going beyond fraud
Shareholder Value
Remote fixes
Early warning
Improved MTTF
Product Analytics
Add on Services
Supplier Quality & Chargeback
Operating Margin Market Position Revenue Growth
March 13, 2015 Proprietary and Confidential - 11 -
Call to Action
Sensor requirements
Sensor selection
Connectivity & data infrastructure
Data Analytics approaches
Base Use case e.g. Fraud
Extend Use case – e.g. RCA
Mechanical Modelling
March 13, 2015 Proprietary and Confidential - 12 -
Reaping the benefits
Ear
ly W
arn
ing
s
Frau
d
Pre
ven
tive
re
solu
tio
n
Qu
alit
y Im
pro
vem
en
t
Inve
stm
en
ts
War
ran
ty C
ost
Indicative Savings
20%
Cu
rre
nt
Co
st
Sta
ck
Imp
rove
d
War
ran
ty C
ost
S
tack
Topline Improvements
Pro
du
ct A
nal
ytic
s
Ad
d o
n
serv
ice
s
March 13, 2015 Proprietary and Confidential - 13 -
THANK YOU!
www.igate.com
For further details, please email us at [email protected]