There is no doubt that the pandemic has forced the industry to hasten its embrace of automation. By the start of this year, over 40% of insurers had either already incorporated some degree of accelerated and automated underwriting into their new business process, or had imminent plans to do so. This was a dramatic increase from 18% the year before — according to research to research.
Pandemic or not, this pivot was always inevitable, and the numbers underline why. Today, over half of the insurers that have already adopted accelerated underwriting report payback periods of less than a year, with another quarter either reporting or expecting payback in just two to three years. Decisions that would once take 20 to 30 days are made in 3, and the huge reduction in the costs of policy-issuance means that these systems rapidly pay for themselves.
Nonetheless, the initial wave we’re currently seeing only scratches the surface of the technology’s revolutionary potential.
Simple vs Deep automation
Much of the concrete activity and investment to date has, understandably, focused on the most straightforward life insurance applications. For example, applications from younger, healthy customers with no underlying health conditions, where risk is low and policy issuance is straightforward. The accelerated ‘decisions’ being made by the software here typically amount to unambiguous ‘if X then always Y’ logic. These are decisions that never truly needed to be seen — individually at least — by human underwriters in the first place.
This simple layer was always going to be the first port of call for automation as insurers started to dip their toes into new systems and ways of working. We all know that digital automation is great at replacing humans with easier, more repetitive tasks.
But this layer is also inherently limited. By definition, this layer can only ever encompass the most risk-free minority of applications. And, ironically enough, the rush for more and more third-party data to feed these systems can actually push the automatic approval percentage down over time. Where the aim is to avoid any need for human involvement from start to finish, for cases with minimal risk, more data can end up meaning more automatic rejections.
One silver lining of the last year, however, is that the need to learn more rapidly has shown insurers the extent to which they can also use data to accelerate decisions in the far more nuanced and complex cases, which would not have been natural candidates for an automated approach before the pandemic.
This is referred to as ‘deep’ automation, where the aim is not to remove the human element entirely, but to create a more collaborative fusion of man and machine, automating as much as possible while ensuring that human underwriting expertise is focused where it is most needed, as efficiently as possible. By making smarter use of data, and treating accelerated underwriting systems as a tool for human underwriters, rather than a replacement, insurers can reap huge benefits in both time and cost across the majority of the applications they underwrite — as opposed to just with one specialty layer.
This is the thinking behind the current industry interest in predictive models and rules-based processing. These systems can deal with far more complex chains of ‘if, then’ logic, drawing on a wide variety of third-party data in sophisticated ways to make much more nuanced judgments about risk without involving human underwriters in every individual decision. The human element is still very much there, but the skills and time of expert underwriters are put to better use on perfecting and tweaking the models themselves, as well as dealing with the most difficult and complex of applications that arise.
While it’s early days and these models are still being developed and improved, the untapped potential is there for the taking with current technology. At the moment, it was estimated that around 10 to 20% of issuances across the industry involve some form of automated underwriting. With properly integrated systems using deep automation, there’s no reason why this number can’t be closer to 70%. And that’s as things stand today, before models and techniques inevitably improve in the imminent future.
Of course, the potential is about more than just industry opportunity. For customers accessing policies that make use of accelerated underwriting, the result is a far more user-friendly experience in terms of speed and convenience. Extending these benefits to more complex cases could herald a new era of convenience and accessibility in life insurance, not just for those in the lowest risk groups, but right across society — including millions with complex, underlying, or changeable health conditions.
Inmediate is an insurtech startup from Singapore that is using the latest technology such as Artificial intelligence, Distributed Ledger, and NLP, making insurance processing and underwriting fast, cheap, and flexible. That gives for better processes, lower costs, improved time to market, and new revenue opportunities.