HOW AI AND MACHINE LEARNING TECHNIQUES ARE CREATING DIFFERENTIATION IN INSURANCE
Alan O’Loughlin and John Beal share some key lessons on the use of data, analytics, artificial intelligence and machine learning techniques
We know that conversion rates of people shopping for home insurance is quite low due to a number of hard to answer questions along the customer journey. The value of prefill solutions to improve the customer journey cannot be overstated.
The latest advances in these solutions may just appear like “lookup on steroids”, but the amount of modelling, linking, artificial intelligence (AI) and machine learning (ML) techniques that go into pulling all that data together and returning accurate information on the person and property is incredible. The result is quicker home insurance quotes and increased conversion rates. Consider the industries – mortgage, healthcare and beyond that could apply these principles and help consumers do more with less.
MAKE SENSE OF THE DATA EXPLOSION
Today an explosive amount of data is collected, but it is vastly under-utilised as many organisations do not have the expertise to bring data together from different parts of the business to create a single customer view. To overcome this problem, we use a unique identifier, LexID®, which together with proprietary linking technology resolves manages and matches information to create one consolidated view of the customer.
AI and ML analytics solutions have changed how insurers use their data to deliver appropriately priced premiums to customers based on their risk. The insurance sector now has access to thousands of attributes (or data points) to differentiate their pricing models.
Now, we can apply machine learning algorithms to identify the most predictive set of attributes – historical cancellation data and gaps in cover through our single customer view being some of the most recent innovations in this field. This helps insurers become more competitive as they refine their pricing strategies and allowing them to write the risk that is right for their book of business.
MAKE SEVERAL SMALLER, LESS VISIBLE CHANGES
The insurance industry does a good job of what we call “thinking big and starting small”. Already the motor insurance market has improved claims processes through virtual claims handling and touchless claims handling. Here, image recognition technology is used to capture damage or invoices, run a system audit and if the claim meets the approved criteria, it is automatically paid without human involvement. This improves efficiency, cuts costs and smooths the customer journey. The next step will be building ‘trust’ scores based on people’s policy and quote history into this process, to help ensure no previous ‘manipulation’ of data has occurred as an extra layer of safety in this automated process.
BROADER USE OF DATA
Within the motor insurance industry, we have been able to use telematics data much more broadly than originally intended. We can use this data to get on the front foot at first notification of loss (FNOL), helping to deliver a better consumer experience post-accident, whilst providing invaluable insights regarding the circumstances of the collision.
In the commercial property insurance arena, AI can provide valuable insights regarding a potential location for a new branch or business relocation – footfall, crime rate, specific times of year or other local circumstances that increase risk. This insight when provided to the customer enables them to take preventative measures if they do go ahead in that location, decreasing risk and loss costs, whilst helping to improve customer experience and retention.
THE IMPORTANCE OF GOOD DATA
As the saying goes, “you get out what you put in”. An organisation can have masses of data, but unless it is cleansed and normalised it can be useless. We do not take for granted knowing who the right ‘John Smith’ is and being able to link a name with the correct address and date of birth.
As usage-based insurance develops, whether through aftermarket telematics devices, smartphone apps, connected vehicles, even in the future from smart home data, all that data needs to be gathered, normalised, standardised so that consumers can enjoy an improved shopping experience based on their needs and preferences.
Image recognition ML techniques gives us the speed limits of UK roads, in real-time. Without this data we could not know with a good degree of confidence that a person may be travelling at twice the speed limit in an urban area. This allows the insurer to make contact with the customer and take the appropriate actions.
We are now also looking at normalising vehicle build data through ML techniques to price drivers’ insurance based on the safety features of their vehicle.
AI AND ML FOR PERSONALISATION
ML algorithms allow us to refine the understanding of risk to allow for differentiation of pricing strategies and more personalised insurance quoting across home, motor and commercial lines of business.
One of the clearest examples of personalisation is in telematics. Driving behaviour data gives us a clearer picture of someone’s driving risk on the road. Drivers then benefit from being judged based on their individual behaviours, rather than paying premiums based on average driving habits.
THE DEMOCRATISATION OF DATA
Perhaps the most important factor in any discussion about data, AI and ML techniques is helping consumers understand how their data is used and how they can benefit personally. Part of this is transparent consent management. Each time a consumer applies for insurance they consent to their data being used to provide the insurer with the best information possible, so they can set an appropriate premium based on the risk. Within insurance, we are focusing more than ever on educating consumers about how their data can be used and evaluated in a way they control and understand.
Again, driving behaviour is a good example where customers can see clear benefits of sharing their real-time data. And finally, another growing example is the data that will become more available with the Internet of Things – like wearable tech. This enables insurers to apply discounts based on healthy and active behaviours.
While there is not yet the AI or ML technology to help us educate consumers about the value and benefits held in their data, we will use AI and ML to automate and process the data consumers are happy to share.
Alan O’Loughlin is director of analytics and statistical modelling, international and John Beal is senior vice president of analytics at LexisNexis® Risk Solutions
In her first blog for The Journal, Anna Barnes explains why not rushing your studies is key to CII qualification success.
In his last blog for The Journal, Oliver Swinburn looks back on a year at the helm and offers some tips for the Christmas period.
Executive summary: Analysis of the operation of non-life insurance in 2018 is based on the compilation of total data (unaudited) from 75 general insurance