This document provides an introduction to the insurance domain, including common areas like life and non-life insurance. It discusses problems in the industry like fraud and increasing rates of fraud. It proposes using data mining techniques like predictive modeling, outlier detection, and social network analysis to develop solutions to address fraud and other issues. Competitors in the text mining space for insurance are identified along with their product offerings and strategies. Target segments and potential use cases for text mining products in areas like agencies, renewals, marketing, and operations are outlined. Fraud detection using social network analysis and CRM for agents are identified as shortlisted use cases.
[2024]Digital Global Overview Report 2024 Meltwater.pdf
Text analytics opportunities in the Insurance domain
1.
2. Insurance Domain – An Intro
Areas - Life and Non-Life (Health, Property, Motor, Accident)
Parties involved – Insurer, Agent, Client
Problems being faced
$70 billion and $230 billion of medical care spending is
fraudulent in the year 2011
Rate of fraud based on exposure to health data was 7 percent in
2009, up from 3 percent in 2008 in spite of existing analytics
Source:HealthDataManagement
Solution using data mining
Predictive modeling
Outlier detection
Social network analysis
3.
4. Category Definition
Quantitative Text Mining
Data Mining In the Insurance Text Mining
In the Insurance Domain
Domain
SAS TAS SAS
IBM SPSS Attensity IBM SPSS
HP Autonomy
5. Competitors
Company Product name Features & Benefits Strategy
Social Analytics
Sentiment Analysis
Strong market share in text mining for
Attensity Insurance Industry Solution Detect early warnings
the Insurance domain
Fraud detection
Customer service
Sentiment Analysis
IBM SPSS Text analytics for surveys Theme Identification
Categorization Market leader in data mining
IBM OmniFind Analytics Intelligent text mining
IBM SPSS Predictive analysis
Clustering
SAS TextMiner
Data importing
Sentiment Analysis
SAS Sentiment Analysis Market leader in domain
Reporting
independent text mining
SAS Ontology Management Ontology Dev and Management
Classification
SAS Enterprise Content Categorisation
Entity extraction
Insurance datamodel
Data management Solution for quantitative data mining in
SAS SAS Insurance Analytics Architecture
Reporting the Insurance domain
BI
Database marketing
Fraud detection
Megaputer Polyanalyst Text mining solution in a niche space
Customer service
Subrogation Prediction
Hearsay Hearsay Social Social Analytics Text mining solution in a niche space
6. Customers (Insurers)
2010-2011 data in India
48 registered life and non-life insurers
Rs.30,000 crores paid up in capital
Rs.3L crores in premium
Source: Insight
2010-2011 data world-wide
$4.3 trillion in premium
Source: Wikipedia
7. Market Segments
Segmentation Criteria
• IT spending capability
• Popularity
• Growth and Operating
Margin
TOP
Country TOP Indian Tie-ups
Insurers
Insurers with
CNP France
LIC India
AXA France Aviva UK
Aviva UK Metlife US
ING Vysya Netherlands
State Farm US Birla Sun Life Canada
Max New York Life US
ING Netherlands Bajaj Allianz UK
Alianz UK Bharti AXA France
ICICI Lombard
AIG US Canada
General
Tata AIG US
8. Target Segments
1. *Europe
2. UK
3. US
* excluding UK
Specific to
Insurance
Attensity
Product
Positioning Quantitative Text Mining
Data Analytics
IBM/SAS
IBM/SAS
Domain
independent
9. Use-Cases in Insurance domain
• Agency force attrition
Agency department • Agent productivity and agent success factors
Renewals department • High lapse in the initial years of the policy
• Identification of customer segment for cross-selling,
Marketing & sales • CRM
department • Analysis of customer needs & behavior
• Information identification/extraction
Operations
• Fraud detection patterns using Text Analytics and
department Social Networks
• Product enhancements
Products department • Market research
• Competition analysis
10. • Augment the product line by focusing on the Insurance domain
• Products that will help customers (Insurers) with large client base
• Products that will optimize operating costs
11. Use-case solved in the market
Risk management
To enhance
product
requirements
Claim analysis
Discover
emerging
patterns
Premium
renewals &
Customer
Retention
12. Selecting Potential Use-Cases
• Agency force attrition
Agency department • Agent productivity and agent success factors
Renewals department • High lapse in the initial years of the policy
• Identification of customer segment for cross-selling,
Marketing & sales • CRM
department • Analysis of customer needs & behavior
• Information identification/extraction
Operations
• Fraud detection patterns using Text Analytics and
department Social Networks link analysis
• Product enhancements
Products department • Market research
• Competition analysis
13. Shortlisted Use-cases
Solution Value by Insurers Technology Requirements
• Sentiment Analysis
Products • Classification
• Entity Extraction
• Rule Engine
Fraud detection using • Mapping Data
Social Network Analytics • Pattern Detection
CRM • Web Crawling
• Ontology Development
• FAQ database
Agents
• User Experience
• Workflows
14. Product VALUE
• Automatic online responses to agents and customers
• Identification of patterns that are responsible for agent
productivity
Features
• Automated warnings about competition, market and customers
• Map customer data sourced from social networks to inhouse
database
• Enhancing agents productivity
• Encouraging agents and customers to feed-in questions to the
Benefits database
• Call center workload reduced
• Track customers for frauds from yet another angle
• Improved customer and agent satisfaction
Value • Better visibility of environment
• Increased bottom-line revenues
15. Pricing Model
• SAAS based subscription or enterprise license products
Pricing discounts when
bundled
Professional services effort for
custom ontology building
Rate plans based on number
of customers the product
satisfies
Data usage based differential
pricing