Controlling Catastrophe Claims via Computer - Insurance Business America
Processing claims after a natural disaster can be difficult, but automation is making the process easier
Catastrophic events are on the rise – and the insurance industry can only keep up for so long. Many insurers continue to rely on historical data to predict their policyholders’ exposure to disasters and, as a result, are not as prepared as they could be to handle mitigation efforts, let alone the claims process. The automation of catastrophe claims could hold the solution to this problem. According to Raj Tumuluri, president of Openstream.ai®, insurers typically have trouble with two parts of the claims cycle after a disaster: promptly communicating with customers and assessing damage in time to earn customer loyalty.
“The state-of-the-art technology can help claim processors a great deal in capturing and/or collecting data and even assessing the extent of damage and the potential cost of repairs, as well as identifying approved vendors for carrying out the repair,” Tumuluri says.
While he adds that the process isn’t usually fully automated except in the case of small claims, it can cut down processing time and effort by as much as 30% to 50% while assisting human experts.
Another notable pain point in the process is that customers typically have difficulty reaching out to insurers’ support teams post-catastrophe, as contact centers are often overwhelmed from the sheer number of customers trying to contact them. To help process disaster claims and customer inquiries, insurers can deploy chatbots, but Tumuluri warns that some chatbots have rigid interfaces and are not proactive, which can turn away customers. Instead, he says, insurers need to employ “embodied virtual assistants” that are capable of goals-based, empathetic engagement with customers. Openstream.ai’s Eva is a “conversational AI” that’s powered by natural language processing, allowing users to interact with insurers’ systems through a self-service experience that feels personalized.
“It is important to make the experience smooth for the customers by limiting the amount of input needed from the customer for claim processing and [having] an adaptive process that requests only what’s absolutely required and gleans the rest from other sources,” Tumuluri says. “AI assistance needs to be proactive in understanding the goals of the caller and be collaborative and empathetic based on the situation.”