Indian Life Insurance Industry and Scope of Big Data, Machine Learning, AI

Just imagine that you are working on a dataset which has approximately 753 crore observations and thousands of features (predictors) that are divided into 195 classes and there is one class say ‘India’ which alone has 137 crore observations.

As a data scientist, you know these classes are imbalanced and the new observation will have higher chances to fall into the class ‘India’ and this class ‘India’ will drive or dominate the entire model.

Yes, when you are building an ML model this will happen and in reality, also, this should only happen.

India can certainly dominate the world of ML, DL and AI, because of opportunities that we have here. Opportunities are everywhere and good thing is that there is a huge talent also available. Remember we have a class of 137 crore observations and this number is enough to dominate the model.

One of the opportunities is in BFSI sector and particularly in Life Insurance sector. AI/ML/DL can play vital role here.

The Indian Life Insurance industry is growing year by year, day by day. Below 2 tables show the growth of various life insurance markets in the world.

The table I.10 talks about new policies issued in FY19 and the number is 2.86 crores. Now just imagine that this is just a number for one financial year and what would be the number if we add historical number as well? So these many customers are there in India who have bought life insurance products. The number is huge may be more than 30 crore. So it means these many observations with hundreds of features. It is a huge dataset to work on.

Source for Table I.5 and i.10 : IRDAI Annual Report FY18-19 https://www.irdai.gov.in/admincms/cms/uploadedfiles/annual%20reports/IRDAI%20English%20Annual%20Report%202018-19.pdf

Here there is a huge scope for the insurance industry to get into field of Data Science and Artificial Intelligence.

Since the number of customers or number of policies to be served is huge, offering manual service to this large customer base is certainly not an efficient way of managing customers.

To rescue, the industry can be helped by Big Data Analysis , ML and AI in various ways.

 Here is my list of areas where Big Data, ML and AI can be helpful.

Prospecting new customers by segmenting them in various baskets

Big data is playing a major role in identifying prospects who can buy LI.

We have various data available of customer right from their buying habits, their financial goals, family history, health conditions etc. Using a ML models insurers can predict consumer behaviors and suggest which customer or which segment is likely to buy insurance. So here insurers can save their efforts in approaching traditionally multiple customer and target the model predictions directly.

Identifying which customers will cancel the policy in Free Look in Period

There is an interesting analysis done by some insurance companies which says that around 7% customer who buys the insurance, request for cancellation either before the issuance f the policy or within free look in cancellation period. This is a loss to company as lots of efforts would have gone into bringing the customer in the pool. For some big insurance companies this amount is in hundreds of crores. ML can help here in predicting such customers and save some amount of the insurance companies.

Identifying which customers will pay renewal premium

This is similar to point mentioned above. ML can also predict those customers who are not going to pay the renewal premium. This is going to be great help to insurers as almost entire life insurance industry is facing the challenge of 13th month persistency and is answerable to the regulator.

Identifying probable mis-sell or Fraud

This is very interesting. ML and AI both can be helpful to identify cases where mis-sell would have been done. There are some coordinates which if matches with some other customer like phone number is matching with some other customer then there might be some chances that the number doesn’t belong to other person and probability of mis-sell is high here as most of the communications goes to customer’s mobile number or email id. ML and AI can help to identify such mismatches and alert insurance companies.

Identifying sales target at organizational and till individual sales person level and probabilities of achieving it

This is always a big headache to any supervisor who is handling a sales team. Setting targets is currently a manual process. ML can help here by offering predictions about the market, segments, number of people who can buy etc and accordingly companies can set their targets.

Servicing existing customers

Here ML can help in terms of setting auto reminders/providing after sales service/offering digitized service like ClickToPay etc.

AI can also help in building tools like Face Recognition to identify the customer/nominee while coming for any kind of claim.

Developing ML based Chatbot/Voicebot to ease customer interaction with the insurance company and get delightful experience.

Big Data, ML can also be useful for important functions like Actuarial (who decides the premium) , Underwriters , Investment Team etc.

There are many functions which play with the numbers like actuarial, investment, Business Intelligence etc. ML and AI will be helpful to them to make their job easy and also increase the accuracy.

ML algorithm can tell an underwriter whether a case to be accepted or rejected basis customer’s medical history, financial history and predicting future as well. This is important as any insurance company’s biggest challenge is allowing an adverse selection. Adverse selections are those customers whose probability of dying is high but this was failed during the scrutiny at the time of login the proposal. This increases the claims.

ML algorithm can also be helpful in predicting future return while investing in any financial instrument.

ML algorithm can help actuaries to decide the premium looking at next 40 years of claim prediction etc.

Apart from these areas, there are many other small and big activities which can be done using ML and AI.

This will not only help insurance company to earn customer’s satisfaction but also help them to achieve newer heights in terms of new business.

So , I see a great future of ML and AI in Life Insurance industry in India. Some top insurers have already started recruiting Data Scientists and AI experts. This is a good sign for those who are aspiring for data scientist career in BFSI. The demand will increase gradually and so the CTC , so prepare yourself for this upcoming sector in ML , AI in India.

Regards

Jayant Rajurkar  

(I am working in life insurance industry for last 16+ years and also perusing a data science course so I have fair amount of awareness of both these domains).

Special Thanks to my colleague Anjali for her inputs