It’s hard to escape the buzz surrounding ChatGPT, an innovative artificial intelligence (AI) tool that uses natural language processing to respond conversationally to human prompts for information and task requests. The chatbot and its underlying generative pre-trained transformer (GPT) language model are taking the world by storm. There is a lot of discussion across various industries — including claims and insurance — about how generative AI might transform the way people interact with technology and, when applied strategically, transform work processes and address key business challenges.
Why GPT is garnering so much attention
While the AI technology behind it isn’t brand-new, some particular aspects of ChatGPT are fueling the current wave of excitement, including:
- The fine-tuning that enables it to follow human instructions and so closely mimic detailed natural language
- Its “training” in massive quantities of information from across the internet
- The ease with which the technology is now accessible to any web user — free of charge and without the need for an application programming interface (API)
Also making headlines are the potential and serious dangers associated with using generative AI outside of secure environments. As outlined in a recent article by Eric Schmitt, Sedgwick’s global chief information security officer, these risks can include:
- Leak, theft or misuse of data
- Lack of control over output content
- Non-compliance with laws and regulations
Considering the challenges and opportunities
Before jumping on the bandwagon with any technology trend, and especially with something as new as GPT, industries and organizations need to carefully weigh the pros and cons of adoption. A general sense that “everyone else is doing it” and it’s the “latest and greatest” technology offering does not justify blindly hopping aboard.
In the claims industry, the risks highlighted above cannot be taken lightly. Given the sensitive and personally identifiable nature of the information we touch every day, claims data must be handled in accordance with strict privacy protocols. (As such, claim details and other HIPAA-protected information should never be used in a publicly accessible GPT prompt; Sedgwick blocks these applications from our networks as an added security measure.) However, with a thorough understanding of the possible risks involved and the types of robust controls that can be put in place to mitigate them, organizations wishing to pursue the benefits of generative AI can safely proceed in integrating such tools into their systems and processes.
At the most basic level, those benefits include greater efficiency. Generative AI can rapidly draw meaningful information from huge amounts of data and automate important but routine pieces of the claims process. More significantly, effectively embedding GPT into claim workflows can give busy examiners and adjusters the opportunity to focus on critical aspects of their jobs that only humans can perform: connect with and care for people. “Outsourcing” to technology some of the mundane and repetitive tasks associated with claims will enhance the work experience of our industry’s talented professionals. With the labor market still tight and today’s workforce seeking engaging and challenging assignments, as well as opportunities to grow and make a meaningful difference, investments in technologies like GPT have the potential to deliver significant, valuable and widespread dividends.
Supercharging our talented professionals
Sedgwick has a longstanding tradition of serving clients’ needs through a combination of skilled and empathetic professionals, cutting-edge technology platforms and tech-enabled solutions. With our primary claims systems already powered by AI technologies, graduating to generative AI was a natural progression for us. But before proceeding with the implementation of a GPT application, we worked closely with Microsoft to ensure all data could remain secure in our Azure environment.
Following a deliberate assessment of the benefits and risks, we recently announced the launch of Sidekick — an industry-first integration of OpenAI’s GPT-4 technology with Sedgwick’s established claims management tools. Sidekick is designed to give our talented teams an advantage in their daily work, so they can quickly gain value from claim information, efficiently relay it to clients and other stakeholders, and dedicate more time and energy to the people whose care is entrusted to them.
As a first step, we’re making Sidekick available to our colleagues within our secure environment for quick summarization and analysis of claim documentation and related data classification. With the support of Sidekick, our colleagues can, with the touch of a button, produce automated summaries of scanned documents and easily add highlights to the appropriate file in our system. They can also use Sidekick to quickly uncover key data to help them complete tasks and make a meaningful impact on claims.
Sidekick is our first use case of GPT, so we’re still exploring its true possibilities for enhancing the claims process. We anticipate that future iterations may be able to produce entire claim summaries, identify risk factors on individual claims and programs, explore emerging data trends, and more. Our hope is that it can help us deliver better quality outcomes, faster resolutions and an improved colleague experience. That’s the kind of disruption everyone can get behind.