Insurers must mind consumer trust before fully embracing automation, writes John Downes
Technology has transformed life insurance underwriting over the last decade, with the next wave of developments set to deliver an even slicker customer journey and more accurate pricing. But, alongside welcoming these benefits, insurers must be mindful of consumer sentiment before they fully embrace technology at the expense of the traditional underwriter.
The underwriting process has already come a long way. One hundred years ago, it was simple, with premiums based purely on age, but over time the underwriting process became more sophisticated. Underwriters began to request doctor's reports and medical examinations, most offices began to retain their own medical officers and reinsurance underwriting manuals became ever more comprehensive. As well as medical history, lifestyle factors such as obesity, smoking and alcohol as well as potentially dangerous hobbies also came under the microscope.
Unfortunately, this requirement for knowledge sometimes led to a cumbersome customer journey. Bombarded with questions, often of a sensitive nature and sometimes requiring further evidence or even a medical examination, some people found themselves waiting a considerable time before they were offered cover. Such a lengthy and labour-intensive process invariably meant that a significant number of potential policyholders dropped out before underwriting was completed.
Thanks to advances in technology insurers were able to introduce more automation into the underwriting process. Underwriting rules engines replaced the human underwriter for the majority of business. Question sets were finetuned, and more sophisticated methods for information gathering, such as tele-interviewing and tele-underwriting were introduced.
Streamlining the process this way significantly improved the customer journey, benefitting advisers and insurers too. As well as lower costs, automation lent itself to the creation of new markets such as the aggregator model.
It's not all positives though. Critics of automation point to the fact that it attempts to make one size fit all, forcing customers into the same pigeonhole. As it's such a process-driven approach, there is a risk that some conditions are dealt with in a manner that lacks sensitivity. For example, being asked stock questions about a mental illness can feel cold and uncaring, so to avoid this, many insurers are continually reviewing the underwriting process to make sure it as sensitive as possible.
Finding the right balance between automated and human underwriting can be tricky. Director at Plan Money, Peter Chadborn welcomes improvement but urges caution: "The underwriting process is the area, above all others, which requires understanding of where improvements have and still need to be made to the protection buying experience.
"A clunky application process can be deeply frustrating for both advisers and customers and, in my experience, is the greatest reason for any low levels of protection selling and protection buying."
The next logical step - are predictive models, which take life insurance underwriting to a new level. By pulling in data from a variety of different sources including medical records, wearable devices and mobile health apps, and using computer algorithms and predictive analytics to identify patterns and trends, accurate rating decisions can be made without any intervention from a human underwriter.
It has some significant benefits. Slicker and faster, an insurer can reach decisions on applications almost without the customer being aware that an underwriting process has taken place. Being able to plug into big data also enables insurers to gain a better understanding of risk: knowing that someone regularly gets more than the requisite 10,000 steps and eight hours sleep a day gives valuable insight into an individual's health risk.
Taking advantage of technology to propel life insurance underwriting into the 21st Century may seem sensible, but consumers can feel uncomfortable sharing health and lifestyle-related information, especially where it's collected automatically. Insurers who opt for automation need to therefore clearly explain the process to both and reassure on any data and privacy questions.
From a regulatory perspective too, insurers will need to be able to demonstrate how they arrived at an underwriting decision, but that won't necessarily detract from consumer suspicions, especially where a predictive model gives an output they don't like.
Changes to the data that can be used, and its availability, can also jeopardise a predictive model. Likewise, the way in which data is used may cause problems with both consumers and regulators, especially where there are concerns about the reliability of the data or unintended biases are created.
These issues are already vexing the New York State Department of Financial Services. It issued an insurance circular letter to insurers in January, raising concerns about the use of external consumer data and information sources in life insurance underwriting.
This stated that it had ‘significant concerns' about the potential negative impact on consumers, insurers and the life insurance marketplace in New York due to the accuracy and reliability of external data sources. Potential issues included a lack of transparency as well as the risk of discrimination for certain classes of individual, with the situation exacerbated by the fact that many external data sources are not subject to regulatory oversight and adequate consumer protection.
Ironing out data issues is one thing, but predictive underwriting also has the potential to rewrite the marketplace. Neither automation nor predictive analytics lend themselves to product complexity, so it might simply reduce choice. Similarly, the concept of mutuality could be destroyed, with more accurate risk analysis even potentially meaning that large groups are excluded from cover.
There is also the overarching question: will automation lead to the end of the line for conventional underwriting and even underwriters? The current mix is around 70% automation with the remainder manual underwriting, but this would narrow further in the move to predictive models. To survive, it may be necessary for underwriters to expand their skills into areas such as data science, using their understanding of life insurance to help shape the new approach.
But, although there are plenty of issues to address, predictive analytics offers significant benefits to all parties. Ensuring it is adopted appropriately and robust governance is in place to secure consumer confidence will deliver improvements in the underwriting process and help to take life insurance to a broader market.
Mr Chadborn concludes: "Efficiency is very welcome but not if it comes at the cost of over-simplicity to the extent that logical thinking is absent. Automated systems work to a point, but underwriting, like claims, has a fair proportion of complex and grey areas. Rather than creating situations where a greater proportion of applications are too readily declined or deferred, we need to work towards covering more people overall."
John Downes is underwriting & claims director for VitalityLife
The funding debate shows no sign of abating
'If we need to know the detail, why not ask the applicant?'
Return to work initiative
A first for the panel