FullscopeRMS Insurance

The Roux Institute and the Institute for Experiential AI Prepare FullscopeRMS to Lead AI for Insurance

Executives at FullscopeRMS recognized the enormous value of artificial intelligence (AI) to enhance their disability insurance products, but they didn’t have the advanced analytics to implement the new technologies. So after a few false starts, last year, the Maine-based Sun Life affiliate turned to the executive director of the Institute for Experiential AI (EAI), Usama Fayyad, to find an AI solution. That partnership became the first of many between EAI and the Roux Institute and companies relying on AI to grow businesses.

“We have plenty of data, but insurance companies are fundamentally cautious entities—that’s what has created their long-term success. If we weren’t, we’d essentially be the investment banks,” says David Messinger, director and associate actuary of FullscopeRMS.

Today though, he says there’s also widespread recognition within the larger insurance community that the technology has changed and it’s time to play catch up: “The data techniques have to be bigger and bolder, and a little bit of efficiency goes a long way for return on investment, particularly when it comes to big money areas like claims.”

While many insurance companies are trying to implement AI in some form, life and health insurers are undoubtedly lagging behind in the adoption of AI. Backing from the data science expertise of EAI and the Roux Institute opened the door for FullscopeRMS, which specializes in disability coverage, to lead the way into this new era of AI for insurance. Through the introduction of advanced analytics tools and AI talent and culture they created a groundbreaking practicum course capable of solving real-world problems and driving business innovation for years to come.

One year into the partnership, the company’s natural language processing model is already diligently scrolling through and flagging inconsistencies between unstructured data and customer disability claims to identify claims of interest. FullscopeRMS has replaced its outdated and painstakingly time-consuming manual process with this new AI solution, saving them precious time and money that’s now better spent investigating red flags. That equates to at least $100,000 in savings, an amount that will exponentially increase—at least 300 fold—as AI is scaled up across the company.

“Other potential projects could equate to very sizable returns on investment,” Messinger says. “It’s not hard when you scale something over the course of a large or mid-sized insurance company to find millions of dollars of value, just by automating a few small steps.”

“The data techniques have to be bigger and bolder, and a little bit of efficiency goes a long way for return on investment, particularly when it comes to big money areas like claims.”

David Messinger, Director and Associate Actuary, FullscopeRMS

A Mutually Beneficial Partnership

As one of the Roux Institute’s founding corporate partners, FullscopeRMS recognized the value of innovating through advanced analytics, machine learning, and AI. Together they created an engagement modeled after a teaching hospital that applies a real business problem to EAI’s professional data scientists. The organizations worked together to enroll students in Northeastern University graduate programs and launched a data visualization and communications course for Sun Life’s actuarial team.

“EAI and the Roux Institute’s data scientists started to understand what our problems were so they could fill in the gaps.”

Gabe House, Head of Disability, Absence & Life, FullscopeRMS

First thing was first: Fayyad and his team immediately sat down with FullscopeRMS to identify and catalog potential AI solutions, paving the way for a future foundation of responsible and ethical AI.

“We built buy-in so everyone understood the possibilities, and EAI and the Roux Institute’s data scientists started to understand what our problems were so they could fill in the gaps,” explains Gabe House, head of disability, absence and life at FullscopeRMS.

After weeks of interviews and industry research, the university’s team of data scientists gained insight into the operations of FullscopeRMS. They identified pain points and manual processes so the company could understand their own AI potential. That work all culminated in an “analytics hothouse,” where representatives from business units could build consensus on priorities and shortlist the most worthwhile projects. In the end, they identified over $30M in savings by potentially building the internal and external talent pipeline. As they honed in on the best AI tool to implement first, Messinger quickly picked up on a crucial piece of advice he’d now give any insurer starting to implement AI into the claims process: Make sure the machine doesn’t make any actual decisions. “Instead, we want the machine to be helping people make better decisions faster,” he says.

The idea behind that first project—using natural language processing to mine for claims inconsistencies—was to build momentum and faith that AI tools will be useful. It didn’t take long for FullscopeRMS to realize their lack of analytics expertise was holding them back. Now, FullscopeRMS leaders are hoping to expand the
partnership with the Roux Institute, capitalizing on the expertise of EAI.

“EAI and the Roux Institute helped us bridge that gap and come up with a solution,” he says.

The Practicum

Once the partnership was established, the Roux Institute and EAI got to work building out an analytics cloud infrastructure and implementing new technology platforms at FullscopeRMS. For that first 12-week project, they landed on a realistic and manageable target that would instill confidence in modern analytics and carry over into bigger, more intensive AI interventions. In a Roux classroom, FullscopeRMS trainees worked with data scientists to automate and scale external text data processing to accelerate the claims investigation process. The model they built uses natural language processing to continuously break down external unstructured data into machine-readable parts so the system and employees can recognize the scope and severity of a disability. It reduces the need for further investigation, allowing Fullscope to identify when a claimant is displaying behaviors that aren’t in keeping with their disability—and it takes it one step further. By flagging additional skills, talents, and interests they can now perform regardless of their limitations, the algorithm directs claimants who can’t perform their previous job tasks to potential new career paths. The project automated operations at FullscopeRMS.

“We can start to really solve our problems.”

Suddenly, business learners, data analytics students, and data scientists were speaking the same language and working toward the same goals of realizing and implementing the power of data analytics. Their work culminated with a presentation by graduate students and corporate learners to the leadership team at FullscopeRMS, demystifying natural language processing tools and the use of unstructured data to strengthen the user experience and future maintenance.

By incorporating the recommendations of the Roux Institute and EAI to its shortlist of future analytics projects, FullscopeRMS walked away understanding not only the value of AI tools, but how to adopt them. And even now, ongoing work by EAI and the Roux Institute is still ensuring the company’s IT environment is fully equipped to read and process that new data.

“Oftentimes, an insurance company might bring in consultants who propose a solution, but that doesn’t teach the company how to recognize its own problems and take action,” says Messinger. “By building talent inside the company and opening up to outside talent, we can start to really solve our problems.”

Messinger is currently working out details for a second course of study with Fayyad and his team to further implement AI tools across the company.

The Big Picture for Big Data

Insurance companies are sitting on endless amounts of Big Data. In underwriting alone, FullscopeRMS has decades of requests for proposals (RFPs) on file. However, the reality is that most are limited by traditional actuarial approaches, like pivot tables and simple averages. Notably, FullscopeRMS isn’t alone in trying to play catch up and learn more multivariate techniques.

“We hope to do everything from extracting policy information from an RFP to reducing the amount of time and energy that it takes to set up a case and give a quote,” says Messinger.

Any single given claim represents an opportunity for any insurance company to understand their customer’s behavior, find the best pricing structure, and more accurately examine risk. Executives learned that AI methods like text analysis and image recognition open the door to insights into claimant motivation. That’s what happens when insurance and data experts work together on the same syllabus under the same roof.

“There are worlds of opportunity. We know they’re out there. The Roux Institute and EAI are helping us get our heads wrapped around all the tools,” says Messinger.