Automation, artificial intelligence and other technologies promise to revolutionise underwriting for both insurers and reinsurers – but they won’t replace the relationships at its heart.
The new challenges of underwriting
The recent Monte Carlo meeting (09/2021) illustrated both the challenges and opportunities facing insurers and reinsurance when it comes to underwriting.
On the one hand, it highlighted the importance of effective underwriting in an era of growing risk and uncertainty. The Covid crisis, high cat losses, and the increasing impact of secondary perils such as wildfire and flood have challenged underwriters in the last few years.
Those challenges continue this year. As rating agency Moody’s noted in its update on the reinsurance industry, natural catastrophe losses from weather events were $42 billion in the first half of 2021. That’s among the highest on record for the period. Hurricane Ida in the US and European flooding since have left the outcome of the year for insurers and reinsurers in the balance. Cat losses for the year could reach $100bn, it warned.
Regardless of how 2021 ends, climate change, emerging risks and social inflation will continue to challenge underwriters. And a prolonged period of low-interest rates will make underwriting performance more critical than ever.
On the other hand, in its form, Monte Carlo was also a vivid illustration of the opportunities for insurers and reinsurers. Held for the second year online, the event underlined the acceleration the pandemic has brought in the adoption of digital technologies.
Again, that’s a longer-term trend and will outlast the virus. Industry events will return in person, but the technology isn’t going away. In fact, it’s still accelerating. To take one indication, in the first half of 2021, insurtechs raised $7.4bn, according to CB Insights – comfortably more than the $7.1bn for the whole of 2020. As a recent report from consultants McKinsey & Company notes, the digital revolution is having a “seismic impact” in the insurance and reinsurance industry.
“Over the next decade, the fully tech-enabled insurer will bear little resemblance to today’s organisation,” it predicts.
Underwriting won’t escape, and the benefits will be felt not just by re/insurers but by insureds and cedents. With the market continuing to harden in many lines and the need for reinsurance and insurance growing, fair, effective and accurate underwriting is in everyone’s interest.
We can take one further lesson from Monte Carlo, however: The opportunities from technology don’t diminish the importance of people. Most attending were thankful to be able to join online but eager to again meet face to face. Relationships remain critical in the industry.
As technology transforms underwriting, people remain at its heart. With risks evolving and becoming more complex, artificial intelligence must complement, not replace, real relationships for insurers and reinsurers. Big data can’t be allowed to push out human-scale interactions.
Fortunately, the technology should allow both to thrive...
Making space for people
Automation, big data, and machine learning can transform underwriting across all types of re/insurance.
These are already proving themselves in primary insurance – driven by customer expectations as much as the drive to increase underwriting efficiency. Millennials, the first generation of “digital natives” to grow up in an online world, represent a significant and, in many lines, untapped market that demands a digital, simpler buying experience.
As Accenture notes in its report on underwriting life business: “Over the years, the process of obtaining life insurance has improved as technology enables a more user-friendly experience. Yet for too many consumers, the process remains arduous with paper applications and medical exams, and a lengthy waiting period to receive a decision. For the 80 million Millennials in the US with a purchasing power of $600 billion, the digital transformation of life insurance can’t come fast enough.”
And again, the pandemic has accelerated existing trends: “Insurers faced with generating revenue in an era where physical contact is limited, must now rely more heavily on health data and evidence sources.”
The technology promises to do more than make underwriting faster and more efficient, however: It promises to make it better.
First, automating capture and evaluation of quantitative underwriting parameters doesn’t just accelerate the process for both re/insurer and insured or cedent but can help eliminate human errors and incomplete data. It ensures all relevant information is captured and captured correctly.
Second, automation and big data analyses enables a greater range of information to be considered in the underwriting decision. Data sets too large for human analysis can now be fed into the process. In some lines, that will even include dynamic data, such as social media and – given permission – health tracking from wearable devices, to take another example from the life business. Meanwhile, machine learning offers opportunities for underwriting rules to evolve dynamically, as analysis identifies correlations that humans may miss.
For insurers, and ultimately reinsurers, this will mean more accurate pricing, which brings benefits for customers and cedants, too: Recognition of risk management efforts and improved differentiation in pricing, for instance. It can also enable insurers to cover risks that they could not otherwise be comfortable with relying on traditional underwriting tools. That will support coverage of not just risks on existing lines but establishing new lines and innovation to support new covers.
Crucially, though, people remain key to the process. First, insurers do not cede control to the software. It is still down to them to determine risk limits, price bands, and the factors to be considered in the underwriting process. In automated underwriting and AI, underwriters are the ghost in the machine. The technology merely streamlines the process and expands their scope.
Second, a central purpose of automation in underwriting is to support triaging. It saves underwriters time in collecting information they would have to chase and enables more straightforward cases to be accepted or rejected automatically (and can set pricing for the latter). In doing so, it frees underwriters to focus on more complex cases.
The aim of AI and automation is not to deny complexity and nuance but to give space for the consideration such complexity requires.
Reading the room
This is as true in reinsurance as in primary business. Automated underwriting and AI will have a vital role in providing coverage required in a challenging future.
Again, this allows companies to draw on new information to make underwriting more efficient and precise: Information that otherwise would be too time-consuming or complex to use.
Recurrent neural network technology, a type of computing system inspired by the workings of animals’ brains, can be used to sift past treaties, for instance. Analysing their context it can determine their structure and the topics covered by the different clauses. Text mining then collects and assesses relevant information and statistics that underwriters can use. AI can also identify correlations that would otherwise be missed.
Combined with automation, for the reinsurer, just as insurers, this has obvious benefits: Most clearly, it removes the time-consuming work studying complex documents that have a range of different formats and structures. It enables reinsurers to manipulate a massive amount of information very quickly. That reduces processing time while improving the accuracy and data feeding into underwriting decisions. As new treaties and outcomes are drawn into the model, the underwriting framework can even evolve.
It can also enhance risk management, identifying key performance indicators in underwriting and claims impacts. That can inform reinsurers to work with cedents and, in turn, their clients, too, to identify risks likely to lead to large losses. It can help move the industry towards greater resilience and help reinsurers and cedents understand the changing risks occurring through developments such as climate change.
And as in primary insurance, it doesn’t replace people but empowers them. This technology enables human-scale interactions and even encourages competition. The technology allows for smaller teams to service books of business that would otherwise require significant scale. Indeed, smaller, newer organisations potentially have an advantage in being able to build the technology and business together, rather than facing the difficulties of incorporating it into existing legacy systems.
It also facilitates relationships. Underwriters are freed from manual, repetitive tasks to spend more time with cedents looking at difficult cases, having the challenging conversations needed, and working with them to insure complex risks and identify the coverage needs of the future.
Reinsurtech will mean fairer, faster and better underwriting decisions. But it doesn’t have to mean an impersonal approach. As the industry begins to meet together again in the coming months, automated underwriting and AI won’t be replacing relationships.
But it will increasingly be informing the conversations.
