Written by Kevin Steer, CEO, One Two One Advisor Sdn. Bhd.
Predicting New Customer Products to Buy
Predicting Product Recommendations for Banks, Insurance, Takaful & Investment requires analysis of existing customers using predictive analytics models to identify the next product they are likely to buy. This also includes tracking of historical loans repayment, investment, fraudulent claims etc.
Once prospects have been identified and assessed for historical transactions; this requires a personalized marketing to that customer, which can be achieved by Robotic Process Automation. RPA can used to create email templates based on particular needs; this can include targeted articles focused on the particular product in mind. RPA enables sending of one or more personalised emails to the customer, for example send article first, then offer upsell product, and notify agent or marketer to follow-up with customer.
Know Your Customer (eKYC)
With the COVID-19 pandemic, remote Customer Onboarding is becoming a priority for Banks, Investment, Insurers & Takaful. However, this has the potential of fraud, thus you need to Know Your Customer, which can be achieved via online eKYC solutions.
Behavioural Biometrics helps Banks, Insurers, Takaful & Investment companies for identity proofing, continuous authentication, potential account takeover fraud, and phishing scams. A customer’s signature, voice, thumbprint, and face are unique as well as how customers interact their device. However, facial recognition is critical, and you do not want someone holding a picture of another face in front of themselves, during onboarding or sign-in authentication. With modern AI solutions, facial recognition can be achieved.
In addition, onboarding also includes capturing the identity card, date of birth, full name, mobile number, and address. These can then be validated with the face.
Chatbots & Robo Advisors
It is predicted that Robo Advisor platform revenues are set to increase to USD 5 billion in 20221
Chatbots have become very popular, offering 24x7 multi-lingual access to prospects and customers, who may want to chat in the evenings when they have finished their work. In many cases, the Call Centres close around 5-6 pm, so are unavailable in the evenings.
Financial customer service chatbots assist customers with financial transactions in a secure environment. Chatbot services include enhancing customer support, promoting cross-selling, reviewing an account, reporting lost cards, making payments, renewing a policy, handling a refund and fraud detection.
Robo Advisors are more intelligent and support more solutions, for example providing financial advice based on a person’s lifestyle goals, understanding their needs and recommending products to address their goals.
Objective of Multi-Lingual Robo Advisors
Benefits of Multi-Lingual Robo Advisors
- Increase Revenue & Reduce Cost by delivering 24x7 Personalized Engagement by providing advice on Lifestyle Financial Goals, FAQs, Products & Emergency services
- Use cognitive robotics to sort & address basic service requests from customers & agents
- Potential to reduce customer service costs by up to 40% by implementing 24×7 intelligent multi-lingual bots
- Communicate with everyone so they can understand
- Provide needs-based advice and recommended products
- Connect with people to generate more leads
- Robo-advisor can also do the important but repetitive task of rebalancing existing customers’ portfolios
Losing customers is a long-standing issue; one of the key issues why they switch are “Unhappy Customers”. This happens in Banking, Insurance, Takaful & Investment sectors; when a customer calls with a complaint, call centres need to be response to address the person’s complaint. Once the complaint is addressed, you need to get customer feedback and track this.
Applying AI to customer churn can be done in a number of ways i) Using Chatbots to respond to customer complaints; ii) Applying Predictive Analytics & Behavioural Intelligence to assess complaints and predict likelihood of churn.
Chatbots can respond much faster than call centres operating 24x7 and can potentially understand the complaint issue. They can then respond and offer advice.
Predictive Analytics & Behavioural Intelligence can assess the complaint and feedback given, matching new customers to previous customers that have switched based on their unhappiness.
Using AI for Insurance & Takaful Underwriting
Insurance & Takaful Underwriting is critical; with the potential for inaccurate proposal data, leading to acceptance of proposals. AI based Underwriting using Predictive Analytics includes statistical data and past history of customers and is used to analyse the risk involved.
An Underwriting score is allocated to the underwriting after analysing the proposal information provided. The predictive underwriting applications score higher using a specified threshold (defined by insurer / takaful) and based on this assessment underwriting follow-up is triggered.
Each proposer has unique underwriting requirements, and is based on a combined risk selection, including medical data provided, age and occupation. Predictive Underwriting uses a number of algorithms to come up with the Risk Score; and then analyses the person’s predicted risk based on a number of parameters. If the predictive underwriting raises a higher score, then there is a need to follow-up.
There are a variety of predictive algorithms, and you can run any number of algorithms to assess which algorithm offers the best results.
Predictive AI for Lapsation & Fraudulent Claims
Predictive Analytics & Machine Learning can be used to assess the potential of customers who may be struggling with payments. Based on the analysis Insures & Takaful can offer a reduced cover at lower premiums to address their needs.
In addition, it can also be applied assess the potential for fraudulent claims; taking into account identification of fraudulent healthcare indicators.
Machine learning & Predictive Analytics is used to understand the nature of the problem, use indicators to identify suspect physicians, and assess validity of the clustering.
Predictive Analytics Objectives
Predictive Analytics Considerations
- Better Predict and Mitigate Risk
- Applies to marketing & sales, customer service, new business underwriting, personalization and claims management
- Estimated that AI can drive cost savings of $390 billion across insurers' front, middle, and back offices by 2030
- Artificial intelligence has the potential to move from a detect-and-repair mindset to a predict-and-prevent philosophy
- Use multiple algorithms to assess the best predictive model
- To ensure that AI in the front office is successful, insurers need to have a clear strategy for implementing AI tech and use it for solution for specific problems
- Use a hybrid model between digital & human to ensure they cater to all consumers
- AI should be viewed as a means to Augment but NOT replace human capabilities
Investment companies are looking at AI to predict future fund growth or decline. A number of companies are now using AI and machine learning for trade execution, investment recommendations, data-driven portfolio creation and risk management.
The fund prediction model applies analysis, looking at the financial health, using indicators such as earnings, price-earnings ratios and dividend yields.
Based on this analysis, then a chart shows whether the price will go up or down over the next five days.
Automated AI Trading systems, analyse real-market data & trader activities, using a dataset to train the predictive model; using a number of complex factors that can affect the stock market price. As more and more data are collected, machine learning algorithms make predictions more accurate.
Robotic Process Automation (RPA)
Robotic Process Automation is having a big impact on FinTech & InsurTech, by automating processes and lowering cost of engagement and support. Robotic Process Automation helps banks and insurers engage customers in real time, increasing efficiency and productivity, and enabling straight-through processing.
RPA robots utilize APIs to trigger processes, capture data and apply applications just like humans do. They trigger responses that communicate with other systems in order to perform a variety of repetitive tasks.
RPA enables banks & finance companies to automate Customer Onboarding, Know your Customer & Anti-money Laundering, Report Generation, Open New Accounts, Loan Application & Processing.
For each different RPA processes, define each type of process and continuously monitor Customer/Prospect status reports.
The benefits for FinTech & InsurTech include:
- Scalability: Enables processes to be automated & scaled, reducing human interaction
- Increased Efficiencies: One RPA processes are setup and validated, operational processes can become faster, and efficient.
- Better accuracy: By following rules, RPA processes can verify data and make decisions.
- Reduce Back-End Effort: Using automation, RPA processes can dramatically improve the speed, by integrating with back-ends and validating data.
- Improve customer service: The RPA processes automate the processing and if there is any follow-up required, the customer service team can be notified & interact with the customer.
- Cost Effective: Enabling a cost saving between 25% - 50% of time and cost
- Risk Compliance: RPA assists in generating Audit Compliance Reports, in order to reduce business risk.
- 24x7 Availability: RPA processes work 24x7, enabling customers to transact anytime.
- Minimize API Integration: RPA systems can use API calls to connect with different modules.
- Implement Faster: RPA services include a UI design of processes, where you can select existing actions. If new actions required, can add more components to the design.
- Use of Customer Data: RPA processes and also check existing legacy data, and use this to validate and automate processes. Legacy Data can also be utilized to generate automated reports.
Steps to successfully implement & deploy RPA in Banking, Insurance, Takaful & Investment:
- Planning: Identify processes which you want to automate.
- Business Impact: There is a need to understand the RPA Processes, and evaluate whether they can be achieved. With this you need to assess if there are existing APIs, or whether you require new APIs.
- Develop a Business Case: Understand the cost to implement and what efficiencies can be achieved. If the cost is too high and efficiencies are minimal, then it may not be worth implementing.
- Develop a Business Plan to Execute: Once the Business Case is approved, develop the business plan for implementation, testing & deployment.
- Support & Maintenance: Once live, thoroughly monitor the RPA process to ensure it works effectively. If there are issues, revise the process and retest.
I believe that Artificial intelligence will fundamentally disrupt & transform Banking, Insurance, Takaful & Investment, helping to reduce Operating Costs, Predict & Mitigate Risk & Proactively Engage with Customers
AI can help FinTech & InsurTech by using
- Chatbots & Robo Advisors to deliver 24x7 personalized engagement to build customer loyalty & advocacy
- Predictive Analytics can help to proactively engage with customers at the right time
- AI based underwriting can help assess risk as well as identifying new products
- Robotic Process Automation can help improve data accuracy & reduce operating costs
- Ensuring security is implemented and analysis of fraud
- However, we need to
- Use AI to move from a detect-and-repair mindset to a predict-and-prevent philosophy
- AI should be viewed as a means to Augment but NOT replace human capabilities
- Clean data - IF data is wrong & not cleansed, then AI can give inaccurate responses
- Source- Fintech Futures: Market Disruption, Leading Innovators & Emerging Opportunities 2017 - 2022, Juniper Research
- Source- insurance underwriting Enhancement in Predictive Model, IJCSET
- Source- Improving Fraud and Abuse Detection in General Physician Claims, ILHPM
- Source- How Will Artificial Intelligence Change the Banking Industry, fintechnews
- Source- The impact of artificial intelligence in the banking sector and how AI is being used in 2020, Business Insider
- Source- Behavioral patterns of long term saving Predictive Analysis,
- Source- Solutions: Forecasting share prices with AI-powered supercomputer, The Edge Malaysia
- Source- Machine Learning for Stock Price Prediction, Azati.ai