Building the future of underwriting

Insurance companies can underwrite more efficiently and cost-effectively using artificial intelligence, creating a substantial benefit by freeing up underwriters to focus on more complex and heterogenous risks, where human judgement is more valuable.

That’s the message that Richard Hartley, chief executive officer of insurtech startup Cytora, will be bringing to Intelligent Insurer’s Intelligent Automation & AI In Insurance event on May 21, 2019.

At the event Hartley will be talking about how commercial insurers can transform their underwriting using artificial intelligence (AI), a part of which is automating repetitive processes which would otherwise be time-consuming and manual.

According to Hartley, Cytora is focused on removing frictional costs in the insurance value chain. He claims that at present for every pound of premium, about 40 percent is lost in the pure cost of delivering insurance, and reducing those costs delivers benefits to both insurers and their customers.

The risk engine

“The Cytora risk engine analyses billions of external data points and combines them with the internal data from insurance companies, it then structures and connects the data algorithmically,” Hartley explained.

“We use these insights to help insurers to select risks better and to set more accurate prices. At an individual risk level, we are bringing more data points into the view of the underwriter, so they can make a more informed decision about whether they want to underwrite a risk, and what price makes sense given the likely cost of a claim.”

From a technological perspective, the most difficult part is structuring and connecting many different data points that have different formats.

The information might include property construction, company revenue, and weather patterns—all of these features need to be structured and connected to build the most comprehensive view of a risk.

Cytora continuously ingests and analyses data from different sources; its risk engine learns the patterns of good and bad risks, building an ever-expanding picture of risks as they change throughout time.

According to Hartley, when a broker comes to an insurance company with a submission, they can  simply enter the name and postcode of the business seeking insurance, and the Cytora API will return all the relevant information needed to underwrite the risk, as well as a prediction of whether a claim is likely.

“That’s the benefit of this data-driven approach—all these data points are being continuously gathered and then factored into a better underwriting decision,” he said.

Data volume booms

“Data production is growing exponentially. One way to think about it is that around 90 percent of the world’s data was created in the last few years,” Hartley said.

“As the world becomes more driven by machines that create data, such as self-driving cars, or transmitting devices, these machines are continuously creating data about companies.

“All of this data can be used in the underwriting process. Insurance companies need to become highly skilled at acquiring it and ultimately using it to make better predictions.”

Hartley noted that artificial intelligence, ie, a machine making decisions, can only outperform human judgement in areas of the market where risks are homogenous and large pools of data are readily available.

“At the smaller end of the market, automated underwriting makes sense. It’s cost effective and it’s more efficient.”

However, he concluded, automation can go only so far. “For larger, more complex risks where you have thousands of different products,  underwriting should be human led.”

“Automation will happen where the risk is simple and very homogenous, and where there is a clear economic incentive for the insurer to automate there.”

Richard Hartley is speaking at Intelligent Automation & AI In Insurance Europe (May 21, London). His talk will focus on how commercial insurers can transform their underwriting using artificial intelligence to automate repetitive processes and remove frictional costs in the value chain.  For more information on the agenda and speaker line-up, visit the website here.