How to navigate the data revolution in underwriting
Insurity’s Adam Harrison on how to harness data to inform the underwriting process.
In the dynamic world of underwriting, one question often arises: How many pieces of data are truly necessary to assess a risk? 10? 50? 100? 250? The answer, as complex and contingent on various factors as it may be, simply boils down to it depends. It depends on the specific product in question, the risk tolerance of the underwriter, and, most significantly, the curiosity of the person assessing the risk.
Historically, underwriters have always possessed an intrinsic drive for data. This thirst for information has evolved over time; from handwritten scrolls to underwriting cards, transitioning to slips and spreadsheets, and finally, modern digital databases. We spend much of our lives looking for new data sources, trying to find a way of exploiting these sources to find that one differentiator, that one piece of detail that will help us understand more about our clients and their assets.
But the ultimate goal remains consistent: how do we leverage this information to gain unique insights into our clients and their assets, seeking that nuanced distinction?
The quest for differentiation through data
While data is invaluable for underwriting professionals, harnessing it effectively has historically been a challenge. Legacy policy administration systems have often proven to be inflexible, producing distorted data outputs that spit out like a crazy cartoon photocopier. But this isn't purely a technological shortcoming.
There’s always been a ‘human vs machine’ conundrum – when striving for good, clean data, you end up locking your system down with validation rules and restrictive workflow experiences. This results in fewer people using it because of its painful user experience. But if validation rules are removed, the data it spits out is rubbish!
AI in underwriting: symbiosis or contradiction?
Don’t worry – AI is the answer. But it isn’t, and nor does it really help with the human conundrum. Many practitioners remain sceptical of AI models, even more so than traditional Policy Administration Systems. Although AI may play a role in underwriters’ decision making, it cannot replace the innate human intuition central to underwriting. It's paramount that we remember the value of underwriter intuition in the decision-making and relationship-building process.
To function at their best, underwriting professionals need to trust the data they use. Human nature and a fear of the new means we need our practitioners to be able to consume the data and build up trust over time. Our job, as technologists, is to show underwriters a vision of what the future holds without the fear-mongering that the AI robots are coming, whilst also ensuring that data presentation remains straightforward and user-friendly.
Incorporating data-driven insights: learning from the automotive sector
Reflect upon the automotive industry, where vehicles continually provide users with essential information, notifying them of potential risks and ensuring the car's well-being – for example, it’s tested the tire pressure, the engine oil, and the temperature and also, “It’s a bit icy out there – be careful”. This seamless interaction instils trust in the driver, all while allowing them to focus on their primary task: driving.
Car manufacturers have figured it out, with a familiar interface that allows you to get on with your activities with the least amount of fuss and fanfare. But if there were to be a problem, ping, here is your alert.
So, why shouldn’t the underwriting experience be the same? Why not have a control panel that lets you do your day-to-day activities with the least amount of fuss, but all the while in the background, the machine is connected to hundreds of data points that are all being checked in real time? This user-friendly interface is not merely about abundant data, but ensuring its presentation is supportive, undistracting, and not overwhelming.
Navigating the data revolution in underwriting
To this end, technology vendors’ primary aim is to construct a simple underwriting interface linked to various data services, constantly surveying these sources for crucial information, but still letting the underwriter tackle the complex task of making risk-based decisions. The primary objective is to seamlessly integrate data into underwriters' daily operations, only drawing their attention when indispensable. At the end of the day, we don’t want the humans to let the machines win – we just want them to be part of the team.
Adam Harrison is enterprise account executive at Insurity and is part of its MGA division, which concentrates primarily on the distribution chain of London market capacity in the North American markets.