More Related Content Similar to Cognitive Computing in Insurance: Going Beyond Predictive Analytics (20) Cognitive Computing in Insurance: Going Beyond Predictive Analytics1. Cognitive Computing in Insurance:
Going Beyond Predictive Analytics
© 2015 IBM Corporation
8 December 2015
Craig Bedell
2. 2
© 2015 International Business Machines Corporation
Cognitive is the new
competitive advantage
for Insurance
3. Watson is a cognitive technology that processes
information more like a human than a computer—by
understanding natural language, generating
hypotheses based on evidence and learning as it goes.
4. Cognitive Computing in Insurance
4
Cognitive
How can we learn dynamically?
Prescriptive
How can we achieve the best
outcome?
Predictive
What could happen?
Descriptive
What has happen?
How can everyone
be more right…
…more often?IBM Analytics
Learning Models
Experience Memory
Feedback mechanisms
Rules
Optimization Models
Recommendations
Predictive Models
Scores
Forecasts
Reports
Dashboards
Visualization
6. …5 key forces are changing the landscape of Insurance
Rising customer
expectations
$ Low interest rate
Sophisticated fraud
Changing demographics
Rapid digitization
Cognitive Computing in Insurance
7. © 2015 International Business Machines Corporation
Cognitive Computing in Insurance
Pushing the current boundaries of:
Innovation
Growth
8. © 2015 International Business Machines Corporation
Cognitive Computing in Insurance
Helps focus on improving:
Engage
Discover
Decide
9. © 2015 International Business Machines Corporation
Cognitive Computing in Insurance
Cognitive systems facilitate
Decision making
Reduce human bias
Scale
Expertise
10. What is Watson ?
© 2015 International Business Machines Corporation
11. © 2015 International Business Machines Corporation
IBM Watson is a technology platform
that uses natural language processing
and machine learning to reveal insights
from large amounts of unstructured data
Cognitive Computing in Insurance
13. Watson is creating a new partnership
between people and computers
that enhances, scales and accelerates
human expertise.
Tabulating
Systems Era
Cognitive
Systems Era
1900
Programmable
Systems Era
1950 2011
14. Corpus
Manuals, FAQs, and
reading lists; i.e. body of
knowledge
Representative Questions
The questions asked to gain
knowledge and competence i.e.
set Ground Truth
Trained Watson
Ready to be put to work
i.e. use knowledge to
learn more and answer
different questions
Watson is Schooled
Watson is taught, not programmed
Cognitive Computing in Insurance
16. What makes Watson different?
Watson understands language. It
understands what is being asked
and is able to find and communicate
answers.
How can Watson be used?
• Engage in conversations
• Explore - find and aggregate
information
• Discover information and
relationships
• Decide - evaluate information
against defined criteria
The Watson Difference
© 2015 International Business Machines Corporation
Cognitive Computing in Insurance
17. Explore - find and aggregate information
Visualize, and discover
insight in unstructured
data through NLP and
text analytics, Content
analytics and cognitive
insights
Explore information from
internal and external
content including
unstructured text, emails,
call center records…
Apply cognitive
capabilities to enhance,
scale, and accelerate
human expertise
Explore
Analyze
Interpret
Cognitive Computing in Insurance
18. Characteristics that represent good uses of Watson:
© 2015 International Business Machines Corporation
High volume of
unstructured data
Users want to explore
multiple answers
English language*
(Spanish & others in
progress)
Support dialogue
when appropriate
Require answers
with evidence
Continuously learn
* multilingual support available
for content analytics
Cognitive Computing in Insurance
19. © 2015 International Business Machines Corporation
Cognitive in Insurance
Cognitive Computing in Insurance
20. Watson Underwriting Advisor
Enable Underwriters to work more efficiently
Read and understand the natural language
contained within unstructured documents,
continuously learn based on feedback
Create a comprehensive evaluation process
to collect, integrate, visualize and
evaluate risk
Use cognitive reasoning to rapidly parse
relevant evidence and generate hypothesis.
© 2015 IBM Corporation
20
Maximum Age Criteria: Born after January 14, 1945.
DOB: 1/15/1950
Total Cholesterol = 276and HDL ratio= 5.1
No information provided in submission
Blood Pressure Reading: Not to exceed 140/80 age ≤ 55, Not to exceed
140/90 age > 55.
BP: 135/85
DUI/DWI history: no incidents in the last 5 years.
Cardiovascular disease: none.
No cardiovascular disease. Last examined 2 years ago.
Universal Life Underwriting Guidelines
Health History
20
Cognitive Computing in Insurance
21. Benefits of the Watson Underwriting Advisor
• Evaluate the submission rather than data
gathering
• Reduces time to process each submission
• Reduction of Underwriting Expenses
• Incorporation of “best practices” expertise for
consistency and accuracy
• Scale underwriting staff
© 2015 IBM Corporation 21
Cognitive Computing in Insurance
22. Transform Underwriting
Make existing decisions smarter by incorporating insights from unstructured data
Underwriter
Review Application
Completeness
Review and Apply
Underwriting Guidance
Recommend
Underwriting Decision
© 2015 IBM Corporation
Summarize & extract
information from
submission
Enrich submissions
by combining internal
and external data
Evaluate products
for performance and
key metrics
Support decision
making with insights
from precedence
Improve decision
making with
continuous learning
Cognitive Underwriting
Structured
Data
Application
Documents
Historical Data
& UW notes
Social Data Loss History
Inspection
Reports
Example Customer Data
Public Medical
Literature
Watson Health
Data
Wikipedia
DBPedia
Public
Financial Data
Historical
Public
Submissions
Weather
Patterns
Example Watson Data
22
Cognitive Computing in Insurance
23. Watson Wealth Advisor
A set of services enabling advisors to understand client needs, engage with
clients based on individual sentiments, personalized events & news and
provide evidence based advice
23© 2015 IBM Corporation
Cognitive Computing in Insurance
24. Self Service Virtual CSR
Extend expert personalized advice.
! Strengthen relationships through better customer
advice & information sharing
! Increase revenue through better informed next-
best-action, cross-sell / up-sell
! Improve operational efficiency
Cognitive Computing in Insurance
25. © 2015 International Business Machines Corporation
Customer Self Service
Insureds
Producers
Field Staff
Cognitive Computing in Insurance
Informed Interactions –
26. © 2015 International Business Machines Corporation
Other Insurance Areas of Focus
Claims Adjusting
Producer Effectiveness
Marketing Optimization & Customer Insight
Legal Risk & Exposure
Policy Language – Legal Interpretations
Cognitive Computing in Insurance
27. Cognitive Computing in Insurance
http://www.ibm.com/analytics/watson-analytics/
IBM Watson Analytics
for Insurance
30. © 2015 International Business Machines Corporation
What will you do
with
IBM WATSON ?
Cognitive Computing in Insurance