Around the globe, organisations of all industries are embarking on digital transformation programmes, creating new business models using modern technological advancements. One such technology that presents a wealth of opportunity to the insurance sector is Artificial Intelligence (AI). Industry analysts Celent describe AI as “technology that makes inferences and decisions that used to require direct human involvement”.
The insurance industry has long been described as slow to embrace change; many organisations still make do with manual processes, tolerate duplication of effort, and resign themselves to clunky practices and reams of paper. Consequently, skilled employees spend their days maintaining the current analogue status quo instead of on value-adding actions.
The sector, therefore, is ripe and ready to embrace the opportunity AI presents. Think about it, automating previously manual human decisions and processes could: 1) enhance the customer experience; 2) improve underwriting and pricing decisions; and 3) speed up the claims process whilst preventing fraud — and reducing cost — mitigating risk and increasing revenue.
This manipulation of technology is already happening. Those that follow the trade media will have heard of Lemonade — a US based insurtech start-up. Lemonade hit the headlines with a trail of publicity claiming they had set a world record of 3 seconds to settle a claim — though I’m not sure this made the Guinness Book of Records! The Lemonade business model uses a combination of machine-learning algorithms based on historical data to automatically manage claims “without touching the sides”. Time will tell when and if Lemonade turn a profit or whether the business is based on clever marketing noise. Regardless; any initiative that delights the customer whilst delivering tangible operational efficiency is destined to raise attention.
One Japanese Life insurer (Fukoku Mutual) is replacing 30 employees with an AI system that can calculate pay-outs to policyholders. The technology reads tens of thousands of medical certificates and factors in the length of hospital stays, medical histories and any surgical procedures before calculating pay-outs. The insurer believes it will increase productivity by 30% and see a return on its investment in less than two years. The firm said it would save about US$1.3 million a year after the US$1.8 million AI system has been installed.
As you would expect, here at SSP we are also embracing AI as part of our solutions for insurers, MGAs, brokers and financial advisers. Our award-winning pricing and data enrichment tool SSP Intelligent Quotes Hub is capable of calling AI and machine-learning tools as part of a quotation flow. This provides scope for dynamically-improved — and more accurate — pricing and risk selection decisions and (maybe by automatically interrogating cold quote data) open up opportunities to price for historically unattractive markets or risks.
We are also innovating across the SSP broker base by using machine learning to offer new revenue-generating potential. Through the use of the new SSP Data Lake, we can run algorithms across anonymised historical data to establish the price point at which a specific risk is bound whilst highlighting those organisations that are willing to offer a price for each quote. We can then apply this learning across our technology to offer brokers the ability to sell quotes where they don’t have the capacity, nor the appetite to price themselves.
As an example: I engage Broker A to obtain a quote for my car but with a string of convictions behind me, Broker A does not want to provide me with cover. Using machine learning it is possible to use my risk data to establish which alternative brokers are highly likely to provide me with a quote and in what price range. Broker A can then choose to offer the risk to Broker B (who does have the appetite to provide a quote in turn), earning a fee for their trouble. What historically would have ended in a “thanks, but no thanks” for Broker A suddenly becomes a new fee-earning opening — and maybe Broker B also earns a commission on a risk that they may not have even seen in the past.
Of course there are data, security and privacy factors to be considered — and brokers need to sign up to the proposition — but it is easy to see the potential in using this technology to grow the business.
It would also be interesting to see whether insurers would welcome a similar business-introducing proposition. In the last few years I have spoken to a number of carriers in the personal lines space who have narrowed their risk appetite to only underwrite the most standard of risks.
But, perhaps again, a proposition such as this would allow them to introduce this non-standard business to another insurer or specialist MGA — earning them new revenues whilst serving their customer or broker better. That said, I recently spoke to one insurer who literally went as white as a sheet at the prospect of their rating algorithms and pricing sweet-spots being open and out there (even anonymised) so we’ll see!
There are a number of other AI tools that insurers and brokers should consider including: Chatbots, Predictive Analytics, Robotic Process Automations, and Virtual Agents. As an example: one large general insurer has employed Amelia as a Virtual Agent to improve their online car insurance quote experience. The overall experience is much more effective and pleasant because Amelia turns the interaction into a conversation, rather than an interrogation, and ensures the customer is always confident about the next step.
The key to choosing the right tools will depend upon the specific business model and strategies of each organisation — as well as having the right specific resources and skillsets to support them (including roles such as data scientists and algorithm engineers). AI has typically been so far outside of most insurers’ IT departments’ area of expertise that it has seemed too risky and difficult. Most insurers simply don’t have the in-house talent necessary to make meaningful progress — and those resources today are still scarce across all industries but this is changing. Google, for example, have introduced their own AI website to attempt to introduce AI and machine-learning to a much wider audience.
What is clear, however, is that those who delay embracing these new possibilities risk getting left behind.
This article is an extract from SSP eye issue 12
About the AuthorMore content by Paul Webster