The phrases “data is the new gold” or “data is the new oil” indicate data has become a commodity in the insurance industry, but how can insurers harness their data to make processes faster and more efficient?

During Carrier Management’s fourth annual InsurTech Summit earlier this month, a panel of three experts shared their perspectives on how data will continue shaping the insurance underwriting process.

“I think the gold really is in using all of that data — and the analytics — to turn it into valuable insights to be able to assess that risk,” said John Willemsen, U.S. head of Distribution at Accelerant, sharing how insurers can use this process to identify niches and competitive advantages.

Changing Data

Jason Wexler, head of insurance underwriting at Procore Insurance Services, explained that while “everybody used to underwrite the same data” on insurance applications, new companies are now exploring third-party data and new data sources that enable them to become more specific and granular in their underwriting.

Previously, for example, underwriters would assess traditional flood maps provided by the National Flood Insurance Program to understand how close a location was to a river and determine risk levels. Now, with better data from satellites, elevations can be assessed. This can significantly shift the risk level of the underwriting assessment, Wexler said.

“It hasn’t moved, the river hasn’t changed, but the data’s telling us a very different story,” he said.

As data is further implemented in the future, Willemsen said he believes underwriters “are going to know more about a class of business and its corresponding lines of business, really, than ever before.” Underwriters’ position in the value chain will become more critical as well, he said.

“Data is absolutely leading the field and enabling some insurers that lean into new data, AI, machine learning and new technology to really accelerate their growth and improve underwriting profits,” Wexler added.

Too Much Data?

Although gathering data is step one in this process, Wexler said, that’s just the beginning.

“Data itself is actually useless,” he said. “It’s really important that we build a model and we benchmark and do comparative analysis to understand a respective or an individual risk and where they fall on the risk profile compared to their group and their peer.”

When asked why some insurers are hesitant to embrace data, Willemsen said it’s not so much hesitancy about the data itself — it’s about the money. Jettisoning successful multimillion-dollar systems built over the years can seem like a hard sell, he said.

“Intuitively, it really doesn’t make sense,” Willemsen said. “A question I would ask everyone is, ‘If you do not … what will happen?’ Carriers like ours are forging ahead into that data-rich environment, and those lagging will likely see a deterioration at some point as a result of that.”

Marcus Daley, technical cofounder of NeuralMetrics, discouraged adopting a master data management approach, which uses a robust and time-consuming process to map data across systems into a structured data source. Carriers moving at a faster pace than the rest, he said, are instead looking at technology like vector databases and transforming it at runtime when they query.

Much has changed since those original, relational databases were created decades ago, Daley added. Instead of trying to mash all of their data together, insurers can separate it and layer interpretative engines on top to produce the same results at real-time speeds.

“The beauty of it is it means if I need to change it — I need to make an adjustment, I want to take a risk and try something new — I can do that literally within hours,” Daley said. “Change a little bit of code … and then I can test it out. That is a totally different way of thinking than what we traditionally think about when we talk about insurance. But to me, that is the path forward.”

Later in the webinar, Daley shared that he sees “an enormous amount of value” in outside data — when it is factual and mathematic. Bias and privacy issues can become concerns.

Wexler pointed to the legal and fiduciary responsibilities insurers have to keep data safe, which presents a unique problem in the industry because data can’t always be deleted quickly.

AI and Underwriting

Wexler said he believes AI can be helpful in organizing structured and unstructured data into formats that make sense for companies, actuaries and underwriters. Already, systems can read insurance applications and look for anomalies between apps and publicly available information on the Internet, such as business websites and Yelp reviews.

Daley said at some point, augmentation and AI will need to handle basic and sophisticated cases, allowing underwriters to move to even more difficult problems.

“I don’t feel like we’re there yet,” he said. “And so, the concern I have is that at some point, a carrier is going to decide to take some risk, and they’re going to do that. And they’re going to be successful. And then they are going to just take the market and move at a pace faster than the rest of the carriers are.”

The full webinar also includes discussions on maintaining a human touch in the underwriting process, why automation can help underwriters focus on underassessed risks and more. Click here to watch a recording of the digital discussion.