Often labelled old-fashioned, archaic and static, insurance is undergoing a dramatic transformation. New players like Amazon, Uber and Lemonade have shaken their respective industries to the core and are leading the way. Differentiated by a vastly improved customer experience and an emphasis on seamless integration of technology, the message is clear: innovate or die.
Insurance may be deserving of its reputation as archaic and old-fashioned; legacy systems continue to prove a headache and many internal processes are still conducted manually. We are often told of the masses of paperwork most underwriters still contend with. And yet, this reputation may become its saving grace; the opportunity to walk untrodden ground is attracting innovation pioneers from all industries.
Chris Castan, Head of AI – Digital Transformation at AXA and Alessandra Chiuderi, Group Head of Analytics at Generali, are no different.
Previously working in AI strategy, Chris Castan was drawn to insurance by the lack of innovative solutions he saw being implemented: “Insurance is well behind”, he begins, when asked what made him pursue a career change.
But where others may have simply seen an innovation desert and stayed well away, Castan saw opportunity. “Insurance has the most potential for AI disruption”, he continues; there are “not many industries that [AI] changes” as much, in terms of benefits for businesses and customers. For those aiming to merge these two facets, insurance appears to be the perfect setting.
Chiuderi agrees. Having previously worked in telecoms, publishing and media, she was always “attracted by the power of analytics on one side and the innovation aimed at improving customer experience” on the other. For them both, “the insurance industry was the ideal opportunity to combine the two aspects”.
Do not re-invent the wheel
On how prepared insurance companies are for wholesale embedding of AI, they were a little more hesitant (as most with substantial knowledge of the subject tend to be). According to Chiuderi, insurance carriers “are still in the process of shaping their digital transformation to be more efficient and provide customers with more flexible and personalised products”.
This requires companies to redefine how they work internally, including re-training or “adding new skills to existing ones”, since very few carriers have the means to hire a new data science or innovation team. Traditional roles must also change. Business teams often ask for the impossible from data scientists, due to lack of technical knowledge. If data scientists can “help the business understand the technology”, business teams can then narrow down their objectives to what is feasible, rather than idealistic. Otherwise, business executives may demand something that isn’t possible, or data scientists may produce a solution that the business does not need.
To become more flexible and not risk resources on doomed initiatives, Castan believes that companies should not be overly ambitious. Business and customer needs, as well as the technology, are changing so rapidly now that solutions can very quickly become redundant. “Have a vision”, yes, but match this “through small, incremental steps”. Those who attempt to “re-invent the wheel” with grand, ambitious projects may quickly find months of work and investment going to waste.
“The technology is there to give good returns”
Many carriers who have invested in AI are yet to see a return on their AI investment. Imperfect data and disparate, siloed legacy systems mean that the data that AI needs to create models and decisions is not readily available. There is work to do here, but also cause for optimism. Castan is admiring when speaking of the new breed of purely digital insurers who have emerged in recent years, particularly in east Asia. Having built their architecture with AI in mind, and employing it for all process automation, the resulting improvement in efficiency has produced “huge ratios”.
So, “when the set-up is there, it works”. But what can insurers do today? While newer methods of data organisation, such as ‘data lakes’, are often discussed in relation to AI, they also address the problem of legacy system data; centralising this data allows for more accurate “automated decision making [which] can save millions.” Chiuderi agrees that “automation is limited by legacy systems”, so provided the required infrastructure is in place, “the technology is there to give good returns”.
“AI makes us more human”
Discussing which business areas would see the biggest impact from AI, both agree that customer experience will see the most significant change for a variety of reasons. For Chiuderi, this rests in AI’s ability to perform time-consuming, “difficult or impossible” tasks. For example, software can analyse speech or thousands of images a second while a claims-handler is on the phone and alert them to any relevant issues.
While this does improve detection of fraudulent or inaccurate claims, the claims-hander can dedicate more time to the claimant, which is essential. As Castan puts it, insurers are “selling the promise that you’ll be there…in times of need” and if “AI makes us [as insurers] more human”, the customer can only benefit.
Whilst other areas such as pricing and underwriting will undoubtedly benefit hugely from advanced, AI-powered analytics, regulatory uncertainty around the use of external data means that neither expert can see past customer experience as the biggest beneficiary of AI. A prominent example of this is the disruptive insurer Lemonade. They are making the biggest noise, says Castan, not because they “changed underwriting” but because their technology-driven customer experience is (or was, at least) unlike any others.
Ask not what AI can do for you, but what you can do for AI
Castan will tackle the subject of AI from a novel angle when he addresses the Insurance AI and Analytics Europe Summit. Reprising his inner John F. Kennedy, he says that for too long, people have asked “what can AI do for me?”, not “what can I do for AI?”. Perhaps inspired by the successes of the new breed of digital insurers, he will explore how insurers can put the correct components in place now, so that AI can be effectively implemented in future. Addressing the scarcity of data-scientists, data architecture, decision-making processes and managing expectations from above, Castan will deliver a comprehensive outline for insurance carriers to lay the groundwork for AI success.
Taking this further, Chiuderi will focus on the strategy behind applied AI, relating to intelligent automation (IA) and customer engagement. While many organizations already employ robotic process automation, there is still uncertainty as to how and where this should be extended into IA. Through an exploration of immediately actionable insights, attendees will gain a detailed understanding of how robotic process automation can be augmented and intelligent automation embedded enterprise-wide. Chiuderi will also touch on how AI can improve customer engagement, using advanced analytics on structured and unstructured data for more accurate customer risk profiling, segmenting and greater product personalization.
Insurance Nexus is holding its 5th annual Insurance AI and Analytics Europe on 9-10 October 2018 in London. Readers of AI News can get an exclusive £200 off their tickets by visiting here and entering the discount code AINews200.
About the author: Gabriel is a content-writer at Insurance Nexus, focusing on the applications of AI and machine learning for insurance. Passionate about politics, the past and music, he also contributes to music and history blogs in his spare time.