Blog(Click here to get to the blog overview page)

What's new? The next wave of insurance automation complemented with new technologies

What was your reaction, the last time you were asked to fill in your details to apply for a new insurance policy? We're pretty sure you grimaced.

Like dinosaurs, "hard copy" insurance application forms are long gone, thank goodness.

I remember when application forms were handwritten by clients, faxed by agents, scanned and then sent to a remote storage warehouse. Underwriters had to dial-in to the company's server to evaluate the applicant's risk. If there was insufficient information to make a judgement call, they had to call the applicant and ask them to fill in more forms.

With the information submitted alongside guidance from the underwriter's manual, the underwriter would need to decide on a premium category, which formed the price of the applicant's protection policy based on their medical knowledge and personal experience. This was subjective and reliant on the underwriter's experience, which is why re/insurers selected a percentage of cases for random monthly audits. The entire process took days, if not weeks, before the agent could provide the customer with a protection proposal. A long way from being an exact science.

How far have we moved the needle?

Fast forward a few decades, underwriting automation means that insurance applicants enjoy a simple, fast, and consistent application process. This computerised, systematic approach provides an end-to-end rule-based system that is simpler for underwriters to use. Now, they can better focus their time and effort in designing tailored questions for the different types of protection policies.

As a result, when purchasing different protection policies, applicants are asked fewer yet more relevant questions. Customers are also instantly notified if additional medical reports are required to support the risk assessment, reducing the turnaround time significantly. The underwriting rule-based system marks the beginning of a customer friendly journey. Re/insurers also receive a more consistent risk assessment.

While the most significant – and beneficial - change occurs on the policy side. Algorithms and data-crunching machines are helping us to write policies more accurately reflecting the client's insured status, and how much he or she pays through the course of the policy. Automation has also meant that we've moved on from a static policy to one that can be dynamically underwritten through its shelf life.

The positive transformation in underwriting automation is also reflected in a recent industry survey, which shows that in the last six years there has been 25% on-year growth in the implementation of automated underwriting engines. However, is this all that the automation of underwriting can offer? No. More exciting is what's to come - automation supplemented with new technologies like machine learning and artificial intelligence, we'll be able to collect and analyse data for more seamless predictive and dynamic underwriting.

Let's take a closer look at how

As we integrate machine learning – a computer algorithm which examines large sets of data and identifies common patterns to solve assigned tasks - to Magnum, our automated underwriting system, we can systematically classify and store new sources of data more effectively and efficiently. At Swiss Re, we recognise the long-term benefits of doing so.

This is the first step towards making predictive and dynamic underwriting more feasible, because both approaches rely heavily on a combination of traditional and new sources of data to create a more personalised insurance experience. By having a standardised approach in sorting and labelling all these new and incoming data sources, it will reduce the time spent categorising data. It will also enable insurers to access data more efficiently in the future, beyond underwriting a policy. We can use the data to track patterns, new and emerging health risk without compromising on security and data protection.

More importantly, with machine learning integrated into Magnum it will radically change the way we analyse data and predict risk during the underwriting process. Let's look at the example of underwriting an applicant who is buying a health insurance policy. When data is put through a logic and an algorithm-based system, we are provided with a generic risk classification such as preferred, standard or sub-standard life….something we are all quite familiar with before. Now when we feed this to Magnum, what we'll get is a precise risk score which indicates where this applicant stands against all others in the portfolio.

This doesn't end here, people with a low risk scores will be provided with a guaranteed or a simplified offer, meaning they will be given a protection policy with fewer questions asked, while others with a higher risk score will go through a full underwriting process with more in-depth health and medical questions posed to them.

As our data analytics capabilities evolves extracting value from vast amount of data, it will enable us to underwrite risk more accurately while providing customers with a more tailored insurance journey. At the same time, all of the above can be done in a more efficient and affordable manner, which also reduces costs to the policy holder.

As Magnum continues to process more data, it will also improve the maturity level of risk underwriting. Hence, by leveraging new technologies, we aim to help close more health protection gaps and create a more individualised insurance experience to meet the unmet customer needs.

Together let's make societies more resilient by being future ready. Now!

Tags

magnum

Author