Rewriting the rules with algorithmic based underwriting
In the early morning hours of 9 November 2016, the majority of U.S. election pollsters woke up to a rude awakening – how did they get it so wrong?
Predictive analytics and masses of data had led virtually all major forecasters and pollsters to predict a comfortable win for Hillary Clinton. What happened? As someone once said, “it’s hard to make predictions, particularly about the future,” a quote brought up by Jim Stanard, Chairman of Ariel Re when speaking about algorithms and data in the latest Voice of Insurance podcast.
Jim Stanard’s views about advanced algorithmic underwriting techniques are refreshing and honest – he doesn’t think there will be a master algorithm that will take over all underwriting decisions and that people will still be involved. As he says, “you can write risks that are not touched by humans, you don’t always need an Underwriter, but it’s a very complicated problem to think about which risks need an Underwriter and which don’t.”
Implied knowledge in underwriting
We think Jim Stanard’s point is about the importance and value of implied knowledge within the underwriting process. Automation and intelligence-based assistance in underwriting do not mean codifying tacit knowledge. The very nature of tacit knowledge, especially in the insurance industry, which is based on years of experience, insight, intuition, and judgements means that it’s difficult to capture and log into a computer system.
Putting aside the complexities involved in underwriting decisions and the amount of regulation in insurance, the underwriting process will always need human interaction at some level.
Navigating today’s uncertain business climate and hard market requires access to accurate data so Insurers can make the right decisions around mitigating risks. The amount of incoming external data that Underwriters need to evaluate is ever-increasing, quite often, some of it being poor data.
Despite digital transformation projects, most of the underwriting process still involves inefficient manually intensive tasks such as data acquisition and data entry, often across disparate systems. A large proportion of an Underwriter’s time is spent on these menial tasks when their time could be better spent on spotting emerging trends, fostering customer relationships, and creating new products.
Intelligent automation and intelligence-based assistance focus on automating those menial tasks, connecting disparate systems, and providing valuable insights from different data sources. Sourcing, digesting, and analysing data, once a manual activity, can be done efficiently and quickly in the background through analytics-based tools and digital assistants for insurers enabling Underwriters to focus on those high-level tasks that generate income.
Automated underwriting techniques and implied knowledge are complementary facets in the underwriting process and should be embraced in unison. Providing contextual information from various data sources and applying an Underwriter’s tacit knowledge on top of that is the future of underwriting. You can read more about this topic in a previous blog we published.
Market challenges and opportunities
In response to the question about challenges and opportunities coming for the industry over the next decade, Jim Stanard had a fascinating response – he sees “two and a half” (that’s correct), fundamental points facing the market. However, we aren’t going to spoil it for you, you’ll have to listen to this very remarkable conversation to find out.