Photo by Yuyeung Lau on Unsplash
Here’s my summary of Session 2 discussed in Day 1 of the First CFA Society India Insurance Conference 2021.
* * *
Session 2: The Implementation of Next Generation Technology in Traditional Insurance businesses, aka Insuretech, with a focus on China.
Speaker: Donald Lacey| Moderated by: Manish Thakur, CFA
Donald Lacey is the CIO of Ping An Global Voyager, a major player in the Insurance business in China. Ping An experienced radical changes in the way it does business as it transitioned from a traditional financial services company into a new-age Insuretech company. The presentation showcased multiple mini-cases of how Technology has enabled incumbent insurance providers such as Ping An to serve better insurance product offerings.
Lacey began his presentation showcasing the dramatic rise of Big Tech companies globally in the financial services business and how they are here to stay. [For more context, read Oliver Wyman’s report on Big Banks, Bigger Techs here. Look at Exhibit 4 if you are short on time.]
This radical take-over was nowhere better illustrated than in China where the likes of Alipay and WeChat Pay now dominate the volumes in small-ticket transactions. In the early 2000s, the slow rise of China’s economic prosperity coincided with tremendous Internet penetration among its citizens. Bootstrapped e-commerce companies such as Alibaba proliferated in a short span of time. But to continue their sustained growth, they could no longer rely on the existing modes of transactions alone. The obvious best source of growth was to integrate electronic transactions with e-commerce. But Credit Card debt wasn't a phenomenon in China unlike in the West. What was interesting however, was the high penetration of the banking system as a majority of citizens stored their wealth in bank accounts.
Their solution- to piggyback off of the deep penetration of the banking system by creating e-wallets that seamlessly allowed transfer of money. E-wallets were a rapid success and their popularity as an alternate mode of payments allowed them to up-sell other financial products over time. Thus began the takeover of tech companies and their growth into Big Techs and Fintechs of the world.
* * *
Ping An’s transition from a traditional business model to that of the Big Techs was based on this foresight of the changing business dynamics. Lacey went on to showcase four significant applications of technology in the traditional insurance business and the resultant benefits:
Case 1.1: Motor Accident claims application and settlement process
Unlike the traditional process, filing for claims has become faster and easier after a motor accident with Ping An. All that a claimant needs to do is open their app, attach personal and vehicle identification documents, and finally record and upload footage of the damaged parts.
Ping An integrated sophisticated Artificial Intelligence (AI) and Machine Learning (ML) based engines in the claims processing divisions to use proprietary Image Recognition technology to (1) verify insured customers and their vehicles, (2) assess the damage and (3) automatically suggest the best course of action- repair or replace. With a high degree of accuracy, the app can assess the claimable value in real-time by mapping the costs of replacement or repairs with their internal database. Using embedded location tracking technology mapped to the database of Ping An’s network of repair stores, the app can also suggest the nearest service store.
Minimising human intervention has saved them monetary and time costs of training human assessors and their travel to the spot of the accident at minimum. Customers are also better satisfied as the AL/ML engines can process claims almost immediately and to a high degree of accuracy any time of the day, month or year. Certainly, any dissatisfied customers can escalate the process to involve human intervention but the benefits of involving technology cannot be overstated. Over time, the cost savings also translate to reduction in underwriting costs and ultimately smaller premiums for such products. That means greater affordability for customers and greater distribution of risk for insurers.
Case 1.2: Livestock insurance claims application and settlement process
Chinese customers who purchase insurance on their livestock live across the breadth and in the remotest corners of the country. Ping An’s proprietary Bovine and Porcine Face Recognition technology has allowed them to process claims in real time just as with the motor accident claims process. The travel times alone were huge cost savings in the traditional business model. Similar benefits are also visible in such products as motor accident insurance.
* * *
Case 2: Agent Performance Tracking
By recording the data generated by Insurance agents conducting grassroots Sales, Ping An was able to manage Human Resources better while simultaneously tracking the status of Leads and estimate Revenues. Mapping the kinds of promotional material shared from the Agent’s WeChat data with the incoming premiums by regions, agents and their managers could track the hit rates of their sales approach and customise their pitches to prospects. Mapping customers by their geo-location and social status also allowed them to upsell or cross-sell other suitable products. Tracking the data on prospects approached with the data on closed sales also allowed managers to motivate low-performing agents and reduce attrition while increasing job satisfaction.
For lukewarm or cold leads, an even subtler technology has proven more powerful. More on that below.
* * *
Case 3: Outbound Voice Calling Robots
By deploying complex heuristics to analyse millions of incoming calls into the company’s call centres, Ping An was able to create outbound voice calling robots that could localise accents while calling prospects in order to close sales. The experiment was actually able to outperform human callers by 2x-4x as machines do not get tired. The technology has the potential to fully remove human involvement as the Voice Calling bots get better at handling more complex conversations regarding information gathering as well as client pitching.
* * *
Case 4: Detecting Fraudulent Activities and tackling Adverse Selection
By taking the insurance application process online, Ping An has been able to filter out applicants that may show a bias for any form of information asymmetry between them and the applicant. They do this by collecting subliminal data that was previously not worth recording or analysing. Using facial recognition data from video recordings of applicants while they are filling insurance documents, Ping An can detect microexpressions and any hesitancy while answering questions regarding their habits, health or other relevant information. Pattern recognition algorithms can recognise any unusual increases in time taken to answer the ‘easy but appearing difficult’ questions.
This application also has the potential to recognise fraud long before it occurs. Using ‘trick questions’ (the phrase being used liberally here), Ping An can also detect if applicants are being dishonest about the details they are filling online. If such signals are evident multiple times, the pattern recognition software can raise potential red flags that would bring attention to the particular case for human decision making.
* * *
All these cases certainly point to an important change in the industry- that the ability to bring analytical process to previously unusable, unstructured form of data into actionable insights will make the business model more efficient and reduce the unnecessary risk borne by the insurers.
If you have read so far, you might also consider the potential ethical implications of gathering and accessing such vast volumes of data from customers. While China gets a bad rep for being callous with consumer data, Lacey was quick to point out that Insurance is a tightly regulated business even in China. Hence, customer data remains protected within the hands of the corporations to which customers voluntarily hand over their data. Such data cannot be shared outside the company for outsourced analysis nor for marketing use.
Lacey also noted the rising demand for skilled workforce that has the ability to analyse big data as well. Ping An alone has an army of PhDs from Data Science, AI and ML realms who are tasked with putting the vast collection of data in their possession to its best use.
* * *
I hope you found this summary useful. To read more of my impressions, head over to my Blog page and search for #IInC21.
Auf Wiedersehen!