The route to automation for ocean marine claims
Is the 350-year-old ocean marine insurance sector ready for automation?
Although a spate of container-lost-at-sea incidents have occurred in the past few months, around 70% of ocean marine cargo claims have less than $10,000 in losses, according to Jose Guerrero, president of Virtual Claims Services, Inc., making the sector ripe for no-touch automation.
“Most marine cargo claims are not complicated as far as data analytics go since there are only four critical documents in claims handling: the bill of lading, commercial invoice, certificate of insurance, and exception receipt,” he told PropertyCasualty360.com. “You can put the commercial invoice on top of another shipping document, such as the bill of lading, and they mirror each other.”
During the 1990s, Guerrero was a senior vice president at Marine Office of America Corp., the largest domestic marine insurer in the U.S., when he noticed most low-value claims required little to no interaction with the insured.
This fortified his confidence that the ocean cargo claims process can be streamlined by targeting low-value and high-frequency claims. He has implemented such systems, albeit with human capital instead of AI by moving low-value claims from senior to junior adjusters. And other marine insurance companies did the same thing.
“Based on the data I was looking at, going back to the late 1990s, it showed that 70% of cargo claims are below $10,000. With these small claims, chances are there will be no coverage issue, and they’ll settle promptly,” Guerrero said. “Why spend human resources on that? Now the question is, ‘how to apply AI to make it end-to-end automation?’ Human resources could then be fully utilized in the handling of complex claims.”
Showcasing the difficulty or perhaps highlighting the opportunity for automating these claims is shipping giant Maersk’s insurance offerings, which Zurich underwrites. The company’s guide to cargo insurance claims states it can take up to two weeks from submission to resolution of a claim.
“If Fokoku Mutual Insurance Company adopted AI in 2017, replacing their department with a digital workforce, and Lemonade recently used their AI engine to handle some of their claims and advertised they achieve three seconds, then why can we not handle low-value and high-frequency marine insurance cargo claims with an AI engine adjudicating claims within an hour? Will the 350-year-old institution of marine insurance quickly adopt no-touch AI?”
Enhanced efficiency & fraud detection
The adoption of no-touch claims operation can address the usual pain points in claims handling, he explained, such as loss adjustment expense, claim cycle, speed of payments or adjudication of claims, productivity, data accuracy and internal personnel issues.
The other advantage of automation would be easier fraud detection. Although new AI cargo applications will have an algorithm to detect fraud by examining datasets and Google location authentication, ocean cargo carriers’ data would be even more beneficial in strengthening fraud detection tools.
This, however, would require a willingness among ocean carriers to share data directly via their “P&I (protection and indemnity) club or independent third-parties with access to AIS (automatic identification system) satellites,” Guerrero said. “Currently, we are working with third parties. Although the data sets from the third parties are not as robust as the operator’s data, the latter is much easier since there is no single central repository of vessel owner’s/operator’s data.”
Another issue at hand is a disconnection between the shipping and insurance industries, Guerrero added.
“Insurance talks to itself, and shipping talks to itself, but there must be some connection. They all have blind spots,” he said. “With the blending of insurance and shipping technologies, there will be good results.”
For instance, every container now has a device to track data. Some go as far as registering internal temperatures, which comes in handy when storing temperature-sensitive goods. However, that data is often not available to the insurers.
“Why isn’t that data available to the industry? When a maritime injury or death occurs, some claims systems of marine insurance facilities have an API connection to the Marine Index Bureau (MIB), which means that when a new P&I claim is entered into the claims system, it is automatically registered with MIB. The adjuster could find out whether the claims is a repeat or not,” he said. “When we eliminate the silos, we will see what I call the single version of the truth.”
Related: