Using the practicum model in business: A new way to find solutions and develop talent

Talent and culture are critical to success in any business, as are the right tools, including advanced analytics. Seeking to grow all three at once, FullscopeRMS partnered with The Roux Institute at Northeastern University to create a groundbreaking practicum course to solve real problems. Our engagement brought FullscopeRMS business learners together with data scientists from Northeastern University’s Institute for Experiential AI (EAI), and Northeastern University students as a project team. The findings from this project are set to drive business innovation in our organization for years to come.

It isn't easy to organically integrate machine learning and AI into an insurance organization from scratch. Data scientists may know the tools, but not understand the deeper business problems. Consultants and vendors can only solve specific problems, and insurance typically isn't considered a cutting-edge space for analytics careers.

Without experience in advanced analytics, machine learning, and AI, business leaders may not fully recognize how much potential exists within their organization. Recognizing this, FullscopeRMS and the Roux Institute created an engagement modeled after a teaching hospital. A real business problem was given to professional data scientists through the university’s Institute for Experiential AI, while graduate students and business leaders worked in a classroom environment to learn how to solve that problem and similar problems in the future.?

The early stage of the engagement focused on helping FullscopeRMS see what possibilities were already out there for us. The university’s team of data scientists gained insight into our operations through weeks of interviews and industry research. We helped them understand our pain points, manual processes, and places where we believed potential for AI existed. The process culminated in an "analytics hothouse," where representatives from all the business units could build consensus on priorities and shortlist the most worthwhile projects to tackle first. Through this process, we identified over $30M in potential savings to our organization, if we could build the internal and external talent pipeline to access it. That is where our learning engagement came in.

Our first engagement focused on augmenting our claims operations analytics with external data, especially unstructured text data. This project was exciting because the techniques for natural language processing were critical to many other projects that we could identify. In our first practicum course, we decided to target something very achievable, hoping to build confidence in modern analytics that we could carry over into bigger but more intensive interventions. Meanwhile, we would be building out our analytics cloud infrastructure and gaining experience with new technology platforms.

Our goal was to automate and scale the external text data processing to accelerate our claims investigation process as much as possible. The basic approach was to leverage the unstructured data, especially with NLP to confirm the scope and severity of a disability, reducing the need for extra investigation. Additionally, the processed information may even reveal talents, interests, and abilities that claimants could turn into new occupations with possible accommodations or retraining. With the engagement, we partially automated the initial research, and the students designed a system to flag important indicators for mental health, activities, and interests that help our claims analysts understand what claimants can and cannot do in daily life.

While graduate students from Northeastern University were gaining practical, real-world experience, they also gaining exposure to the space of opportunities and culture that FullscopeRMS and Sun Life had available to them. Meanwhile, corporate learners from Sun Life and FullscopeRMS, who were also students in the class, were able to see how powerful analytics could be leveraged within our work, not just on social media, but on a host of similar problems that are currently done manually at higher costs, or not even attempted today.

At the end of the 12-week classroom engagement, the students and corporate learners presented their solutions to our business leadership team. While gaps in our infrastructure prevented us from immediately deploying these analytics directly into production, this project introduced automation solutions that will help us collect and process external data in the future. The tool they built considered not just the analytics, but the problem to be solved, the user experience, and future maintenance needs.

In addition, we also added several of their recommendations to our shortlist of future analytics projects. These were ideas that hadn't occurred to anyone during the initial hothouse, but now seemed completely achievable through our new experience. For our teams, this also resulted in a de-mystification of using NLP and unstructured data as part of our now expanded repertoire of AI analytics tools.

FullscopeRMS is always seeking innovative ways to increase our ability to make the right decisions faster and with better outcomes for everyone involved. We understood that predictive analytics, machine learning, natural language processing, and artificial intelligence could help shape our future, but building the experience, talent, and tools for adoption was a daunting task. By partnering with the Roux Institute, we found we could kickstart that process in an innovative way, and the benefits are only just beginning to be realized.

FullscopeRMS is the brand name for services offered through Disability Reinsurance Management Services, Inc. ? FullscopeRMS 2022