AI Jim, the chatbot that handles claims for InsurTech Lemonade, is renowned for its ability to handle claims in seconds.

Two seconds to be exact.

That was the time it took the chatbot to settle a bicycle theft claim in June last year, breaking AI Jim’s previous record of three seconds.

But when Chief Claims Officer Sean Burgess recently spoke to CM Deputy Editor Elizabeth Blosfield about the benefits that the use of AI brings to the property/casualty insurance claims process, Burgess highlighted AI benefits beyond speed.

“What [the AI] really does is it allows the right claim to get to the right adjuster,” Burgess said. While instant claims services from a chatbot is exciting to talk about, improved workflow management using AI is something that resonates with the veteran claims professional.

Burgess, who spent 28 years at USAA before signing on to lead the claims function at Lemonade, described how workflow management, fraud detection and efficiency benefits from AI are helping Lemonade to lower its loss and loss expense ratios and deliver good service to customers on their worst days—when they have claims—during the opening session of Carrier Management’s InsurTech Summit 2024 in mid-May.

Related: The InsurTech Summit 2024 by Carrier Management

“That philosophy of workflow management is not new. Lemonade is just upping the game with AI to make it happen,” Burgess said,

“Getting the right work to the right people makes a big difference. That matters. And by the way, AI Jim is one of those people,” Burgess said.

“If we can get the right claims to AI Jim that he can process through, sometimes in seconds, that’s a beautiful thing. But at the same time, the same technology and automation and algorithms that are running behind the scenes also make sure that the simple claims, which do need a human touch, go to our newer employees [with] a certain skill set, and the more complex losses go to our very seasoned and technical experts,” said Burgess, emphasizing that accuracy is key to improving a carrier’s loss ratio.

“The human element is just amplified by the artificial intelligence that starts right at the front,” he said, referring to Lemonade’s team of human adjusters, referred to by the company as advocates.

According to Burgess, AI is improving Lemonade’s claims process right from the beginning, with AI Jim taking 98 percent of the company’s FNOL (first notice of loss reports), making the report process seamless and easy for our customers. “In fact, over half of those, then, are settled instantly,” Burgess said.

“Are we trying to handle every single claim instantly? Is everything going to be touchless? The answer is no.”

Sean Burgess, Lemonade

Next, AI is working behind the scenes to detect fraudulent claims. “The majority of people are filing a claim for the right reasons, and they’re getting reimbursed for things that they should,” he said, noting that AI can gather data, organize data, and then form an initial opinion to flag the few that claimants trying to take advantage of the system. “That process in itself is not is not new. But the AI doing it is what’s new and is—and it allows for things like instant claims and other things to happen because you have those fraud algorithms running behind the scenes,” he said.

Burgess noted that Lemonade’s AI models get smarter every single day, and fraud detection becomes more precise—helping to move the majority of claimants that “don’t need an extra investigation or an extra conversation” right through, while funneling the problems out.

The carrier is focused on “making things as easy and comforting—and fast—[as possible] for customers that have had something bad happen…

“You have a claim. It’s unexpected. You need someone to help you. That’s what insurance is there for. And that’s what Lemonade is for,” he stated.

“The intent is not to break the record. We’re just trying to get better at what we do for our customers,” Burgess said later, explaining the three-second settlement, which involved a customer talking with AI Jim chatbot through Lemonade’s app, and AI gathering information like police reports and running algorithms in the background.

A claim, he said, is something that’s not expected. “It’s disruptive, and in many cases it’s your first time” filing a claim or that particular type of claim, he said, offering a customer perspective. “How do you take the periods of uncertainty, ….the fear of the unknown and just reduce or eliminate that? That’s what we focused on,” he said, reiterating that AI Jim removes that period of uncertainty on its own in half of the claims intake situations.

“You’re good. We’re here to take care of you” is the quick message to the customer, he said, contrasting the slower process that’s common in the industry. After the FNOL, the claims “sit on a somebody’s work queue. [Customers] wait for a phone call or an email 24 hours later. [That] period of uncertainty is what I’ve learned over my career is what creates the angst right, and the customer service deterioration.”

Burgess also referenced an internal AI tool called Blender that human advocates use to stay on top of those claims that aren’t handled by the AI Jim chatbot.

“Are we trying to handle every single claim instantly? Is everything going to be touchless? The answer is no,” he said, distinguishing a bike theft from a home fire. “If you have the most unfortunate event where your home burns down, we’re going to be there for you. We’re going to call you immediately. We’re going be there with you from the start of that process to the very end because that’s not an instant experience. But throughout that process, the AI will help us be more efficient,” he said.

Burgess also responded to questions about challenges that Lemonade faces in using AI, at one point explaining how Lemonade handles the specific challenge of making sure its models remain free from bias. He also offered advice to insurers who are just getting started with AI on how to best proceed.

View the AI Claims session on the The InsurTech Summit 2024 website to hear more on those topics.

During the AI Claim session of the InsurTech Summit, Blosfield also spoke to Frank Giaoui, founder and CEO of Optimalex Solutions, and Shane Riedman, general manager and vice president of anti-fraud analytics at Verisk.

Like Burgess, but for different reasons, Riedman, who manages Verisk’s claim search platform connecting claims and anti-fraud professionals across the United States, said that AI technologies should not automate every insurance claims transaction. “Some elements of an insurance transaction still require reasoning and judgment—or even morality. And those things do not exist in any AI technology today. The technology is not capable of morality, for example,” Riedman said.

“We are adding science to the art of claim adjusting. We are not replacing the art,” said Giaoui, whose company uses AI to predict casualty claims outcomes and provides AI claims workflow tools to casualty insurers and attorneys. “Once there is confidence and trust in AI, some part of the processes will be fully automated, [but] some will never be fully automated,” Giaoui said, drawing an analogy to airplanes with automatic pilot mechanisms. “You have had automatic pilot available for the last 30 years, but you still have two pilots in any airlines…It will be very similar in in claims,” he said.

Giaoui went on to provide quantification of the accuracy and efficiency benefits that all the speakers discussed, reporting a 25 percent productivity benefit achieved in three years’ time, and indemnity payouts reduced by 3-5 percent for day-to-day claims and up to 14 percent for large verdicts, as well as a 10 percent drop in legal spend.

Read more about Riedman and Giauoi’s Summit remarks, including Riedman’s take on staying ahead of bad actors using AI to submit fraudulent claims, in the related article, “How AI Can Reshape Claims Management.”

Read more about Optimalex in the article, “InsurTech Profile: Optimalex’s AI Tools Co-Pilot Alongside Human Expertise“