Aviva: Secret
HowAIistransformingInsurance
KarthikanSelvarajSeniorDataScientist,
QuantumAsia,AvivaLtd,Singapore
Aviva: Internal
Artificial Intelligence is out to steal your jobs and kill you…
Aviva: Internal
Not really, but Artificial Intelligence is transforming businesses across the world.
Aviva: Internal
And the art of the possible is constantly evolving.
Linear Regression GLMs Random
Forests Graph
Analytics Transformers Word2Vec
ARIMA Risk Modelling
Isolation Forests
Multi-Layer Perceptron
Generative Adversarial Networks
LSTM
Statistical Modelling Decision Trees Support Vector
Machines Deep Neural
Networks Auto encoders BERT
One-way Analysis
Bayesian Statistics Perceptron
Convolutional Neural
Networks VGGNet ELMo
Logistic Regression Markov Chains XGBoost
Recurrent Neural
Networks AlexNet GPT-2
Aviva: Internal
But first, what is AI?
“ifit’swritteninPython,it’smachinelearning.Ifit’swritteninPowerPoint,it’sprobablyAI.”-awiseman
Aviva: Internal
But seriously, what is AI?
“AIreferstomachinesthatcanlearn,reason,andactforthemselves.Theycanmaketheirowndecisionswhenfacedwithnewsituations,inthesamewaythathumansandanimalscan.”
Source:https://www.technologyreview.com/s/612404/is-this-ai-we-drew-you-a-flowchart-to-work-it-out/
foraparticular(well-defined)task
Aviva: Internal
AI has come a long way and has made massive breakthrough in recent years.
Source:https://qbi.uq.edu.au/brain/intelligent-machines/history-artificial-intelligence
1950: Alan Turing’s Turing Test 1955: ‘AI’ got coined 1997: Deepblue beats Kasparov 2011: IBM Watson beats Jeopardy! 2016: AlphaGo beats Lee Sedol 4-1
Aviva: Internal
This AI flow chart from MIT will help.
Source:https://www.technologyreview.com/s/612404/is-this-ai-we-drew-you-a-flowchart-to-work-it-out/
Canit‘See’?Canit‘Hear’?Canit‘Read’?Canit‘Move’?Canit‘Reason’?
Aviva: Internal
Can it See?
Aviva: Internal
Can it See?
Aviva: Internal
Can it See?
Aviva: Internal
Can it See?
Aviva: Internal
Can it Hear?
Aviva: Internal
Can it Read?
Aviva: Internal
Can it Read?
Aviva: Internal
Can it Move?
Aviva: Internal
Can it Reason?
Aviva: Internal
Can it Reason?
Aviva: Internal
Can it Reason?
Aviva: Internal
We can use these techniques to unlock value across the insurance value chain.
Unlock value
Automated smart underwriting for life protection
Machine-learning for GI pricing
High risk advisor detection
Advisor attrition prediction and prevention
Aviva: Internal
AI-powered Smart Medical underwriter for Life Protection
Can we reduce underwriting effort and long turnaround time for Life
insurance applications with medical disclosures and attachments?
us
our
Aviva: Internal
AI-powered Smart Medical underwriter for Life Protection
Can we reduce underwriting effort and long turnaround time for Life insurance applications with medical disclosures and attachments?
Aviva: Internal
AI-powered Smart Medical underwriter for Life Protection
Can we reduce underwriting effort and long turnaround time for Life insurance applications with medical disclosures and attachments?
Aviva: Internal
Machine Learning for GI Pricing
Can we improve loss ratios by leveraging on AI to enable better risk segmentation through price?
y = f(x)
Pure risk cost
Rating variables Predictive model
Aviva: Internal
Machine Learning for GI Pricing
Can we improve loss ratios by leveraging on AI to enable better risk segmentation through price?
Predictive strength of Rating variables from a gradient boosted algorithm
Aviva: Internal
Machine Learning for GI Pricing
Can we improve loss ratios by leveraging on AI to enable better risk segmentation through price?
Partial dependence by driving experience
Partial dependence by NCD
Aviva: Internal
High Risk Advisor Detection
Can we tap on advisors’ production, demographic and customer demographic data to identify high risk selling behaviour?
Aviva: Internal
Advisor attrition prediction and prevention
Are we able to reduce advisor attrition by having a forward looking view on possible attritions through smarter targeting of retention policies?
Aviva: Internal
Advisor attrition prediction and prevention
Are we able to reduce advisor attrition by having a forward looking view on possible attritions through smarter targeting of retention policies?
#thefutureisnow#artificialintelligence