The elixir of business in the digital age is a seamless customer experience. With life insurance, that experience goes well beyond the purchase, demanding – quite literally – a lifetime of engagement from consideration to claim.
Recent advances in artificial intelligence indicate the potential to transform each stage of the complex insurance customer journey, radically altering the entire intermediation process of buying, and living with, an insurance policy.
AI and insurance today
Many insurers may already employ artificial intelligence, without necessarily referring to it as such. It often summons images of IPsoft’s cognitive agent, the human-like Amelia, but it’s not always a smiling, blonde robot. In fact, many insurance firms already rely on less obvious AI techniques, such as predictive analytics, or the decision trees and algorithms of automated underwriting. Given the various forms of AI, and the huge scope of applications, it’s worth defining the practice – broadly, but clearly – as machine simulation of human decisions and actions.
Fundamentally, it is the power of scale of computer-driven intelligence that will force the industry to go even further with the technology to reshape the way insurance is provided and experienced. Lemonade, the P2P platform that has just launched in New York, certainly promises a radically different customer experience. Maya, its (blond, smiling) cognitive agent, communicates with customers via mobile – it takes just 90 seconds to obtain coverage and 3 minutes to be paid for a claim. Lemonade’s tagline, “instant everything”, seems fairly apt. For the traditional insurer lacking the hot-pink branding and novelty of Lemonade, however, artificial intelligence adoption is just as urgent, considering its capacity to pioneer new pathways to – and beyond – purchase.
AI and insurance tomorrow
Consideration and selection
Imagine the ability to speak directly to your phone, or type – in natural phraseology – and have your insurance queries answered immediately.
Recent breakthroughs in Natural Language Processing, coupled with the implications of voice and facial recognition technology, could mean the consideration-to-selection process is completed in a matter of minutes between the customer and the mobile device in their hand.
RBS and NatWest have recently announced that December will see the roll-out of Luvo, a virtual agent designed using IBM Watson technology, with the aim of improving customer experience without employing extra manpower. The chatbot has deep learning capabilities meaning it will learn from every human interaction, and apply its ever-growing knowledge to future exchanges. Luvo even has the emotional intelligence to decipher whether the customer is happy or frustrated, and will adjust his tone accordingly.
Application & underwriting
At the application stage, automated underwriting has the capacity to positively alter customer perceptions of the industry. The speed and efficiency of the process will reinstate life insurance as a priority once again, able to match customer expectations and experiences as shaped by other industries. With the constant support of virtual assistance operating in natural language, unintentional non-disclosure could well be minimised, ultimately resulting in higher completion rates.
Artificial intelligence will not replace human advisers, but engage them more meaningfully. Indeed, Ernst & Young have warned against over-automation and the loss of human insight and perspective, arguing that “models cannot and should not be the final or absolute arbiters in underwriting decisions”.
By offloading the mundane tasks to automated processes, however, advisers will be able to support customers on more complex issues, and IBM reckons they could increase their client workload by “one or two orders of magnitude”. Streamlined underwriting engines could therefore contribute to financial inclusion by enabling access to unserved or underserved customer segments.
Monitoring and claims
Beyond the point of purchase, more powerful technologies will enable smart policy tailoring by constantly scanning data to assess whether the customer would benefit from a different offer. Based on vast customer knowledge, an intelligent machine could enable more efficient, tailored bundling. Such granular profiling means customer behaviour will not just be understood, but anticipated and pre-empted.
And when it comes to the point of claim, intelligent algorithms could be used to estimate the possibility of fraud. With 80% of claims information being made up of unstructured data, text analytics could reduce the cost and time of claims management drastically. Thanks to the advances made in monitoring social media’s constantly updated content, there could be a day when fraudsters are routinely caught out by information mined from Facebook.
Altering customer perceptions
At a time when insurers are struggling more than ever before to communicate the value of protection, artificial intelligence would enable the offer of tailored pricing and truly personal propositions.
The various stages of the customer purchase journey are likely to change slowly and unevenly, subject to the technology itself and to the industry’s appetite for innovation.
But the ideal of decision-free, individualised insurance seems increasingly tangible. Rather than merely being customer-centric, the industry may truly be able to move towards what PwC refers to as the “segment of one.”
With the proliferation of data from myriad sources – wearable technology being but one – and with the algorithms responsible for assessing that data maturing, AI seems poised to help insurers explore how to cater to the quantified self.