Machine Learning Adoption: Crossing the chasm for banking and insurance sector
1. Machine Learning Adoption:
Crossing the Chasm for banking
and insurance sector
Rudradeb Mitra | http://www.linkedin.com/in/mitrar/
EFMA Operational Excellence Council
23rd March 2018, Athens, Greece
3. Content (~30-40 mins)
• Bio
• Why to use Machine Learning?
• What problems to solve using Machine Learning?
• How to solve those problems?
• What are the Challenges?
4. Brief Bio
• 2003-2009: Worked in European Research Labs, Universities and
Startups on AI/ML. NLP, Semantic web, Declarative Language.
• 2009-2010: Graduation from University of Cambridge.
• 2010-2017: Built 6 startups in 4 countries. One of them an ML startup in
Belgium on profiling risky drivers.
• Mentor of Google Launchpad, Writer and Speaker.
5. Objective
• To show you something that will add a lot of value and
you can start working in a week.
12. @copyright: Rudradeb Mitra
Technology adoption
Full article: https://www.linkedin.com/pulse/crossing-chasm-stop-identifying-cats-start-creating-value-mitra/
13. Focus Area
Operational
Excellence
Machine
Learning
Increase transparency and frequency in
communication
Refocus on client risk/return profile
Focus on easy-to-access and easy-to-deal
with private bankers and simplify process
Enhance pro-activeness and anticipation of
client needs
Simplify certain products
Reduce # of relationships per Relationship
Manager
Enhance array of products
Enhance technical and product skill set of
private bankers
Produce more reliable, timely management
information and comprehensive external
reporting
High Medium LowRelevance
14. What problems to solve?
• Are there problems where Bayes Error rate is
>80% and which have high cost?
15. • Solving problems that were thought unsolvable (For ex,
Anticipation of clients needs, Loans)
• Solving problems that were thought not a problem (For
ex, customer acquisition, retention)
• Improving upon existing systems (For ex, Increase
transparency and frequency of communication)
Three groups of problems
18. 1. Identify people who say they need the product
2. Identify people who may need the product
3. Identify people who will need the product
Customer acquisition
19. Identify people who will need the product
The only way to sales conversation with
high-value prospects is to interrupt them
- Fanatical Prospecting, Jeb Blount
20. Identify patterns in your best customers
• Who are your best customer
• Why they became customers
• Why they still buy from you
• Why do prospects choose you over other similar products
38. Challenges
• Technology: Very low
• Data: Yes esp. lack of data. Find innovative ways.
'A simple model with more data is accurate than a complex
model. Most ML system do not need PhD'
39. Challenges
• Adoption: Yes but follow the points mentioned.
• Cost: High but
• one does not need Phd in Data Science. Get the right person
for the right role, even students.
• Get external person for short term project to architect and
build.
• Cost can be quite low!
40. Opportunity cost of not doing!
• Amazon became a bank!
• New 'banks' are already in Asia (India, Vietnam).
• Existential crisis in era of Internet and Globalization.
41. Machine Learning is NOT a Technology Challenge
Adoption
How to collect data?
Intuitive
Thinking
Feel free to contact:
https://www.linkedin.com/in/mitrar/
mitra.rudradeb@gmail.com
Challenge is in
Algorithm
How to deal with
incomplete data?
What data to collect?