When COVID-19 first struck, insurance companies around the world had to scramble to figure out their exposure to the systemic socio-economic event. No-one had a clear idea about the impact of the pandemic, and so the first problem they had to solve was to work out how many contracts in their portfolios could potentially pay out for pandemic-related claims. For large, incumbent insurers, doing that analysis on a broad portfolio of policies was a huge challenge due to the urgency of the task and the strict compliance requirements surrounding it.
To deal with this, some insurers created COVID taskforces, who were asked to read their entire set of policies and find out if there were explicit exclusions related to pandemic, communicable disease, or virus. For humans, such a task is cumbersome and prone to error; but for technology – specifically, artificial intelligence (AI) and natural language processing – such a task is a breeze.
Throughout the pandemic, insurance companies worldwide have started to understand the need to standardise contracts and remove grey areas around unexpected phenomenon, like COVID-19. They’ve realised the benefits of AI and natural language processing in translating and standardising unstructured documents.
According to expert.ai, a global provider of such solutions for the insurance industry, more than 80% of the industry’s entire value chain is represented by unstructured documents, including claims, policies and contracts. In augmenting the capacity of the human workforce by having technology analyse documents and structure data in a uniform manner, insurers can speed up their processes and reduce their risk of errors and omissions.
“One of the typical uses of AI and natural language processing is to structure insurance documents in order to gain a better understanding of the exposure within an insurance portfolio,” said Pamela Negosanti (pictured), global VP of insurance at expert.ai. “Once insurers have structured the documents within their portfolio, when there’s a new event like COVID-19, they know what exposure they have, or they can easily find out what exposure they have by grouping together documents that have an explicit [virus] exclusion, and [those that don’t]. Then, in the case of another Black Swan event, they don’t have to go back and look manually at every single policy.
“That’s what happened at the beginning of the COVID-19 pandemic. Insurers asked teams of people to check through their portfolio of policies to find out if there were exclusions related to ‘virus’. But then they realised that ‘virus’ also related to cyber risk, so they had to ensure the ‘virus’ they were looking for was related to the pandemic and communicable disease. You can’t just do a keyword search for this sort of thing because there’s some ambiguity. This is where AI can help because it has the capability to analyse the context and not just find the word. Doing a keyword search is not enough because insurance policies can be complicated, and they have to be understood in their entirety.”
A second key benefit of incorporating AI is speed. Insurers can use AI tools to speed up their processes, especially when it comes to claims processing. In terms of COVID-related claims, insurers have received a huge amount of communication. While some claims may not be valid and won’t be reimbursed, insurers still have to go through each and every document.
“AI can be used in the front- and back-office to optimise operations, to standardise processes, and offer a better customer experience,” Negosanti told Insurance Business. “These types of technologies can be used to guarantee reduced processing time. Customers want to receive an answer to their claim in two days rather than two weeks, and they’re happy to know insurers have an objective way to evaluate their claim [with the assistance of technology]. By using these tools, insurers can improve the customer experience, and therefore, improve customer retention.”
