For benefits, there's more to artificial intelligence than just chat

Machine learning can spot emerging health trends and build better benefit packages.

If you’re even the slightest bit curious about technology, chances are that you or a colleague has experimented with one of the artificial intelligence platforms that have captured so much attention. Perhaps you had one summarize a long, boring policy manual. Or maybe a coworker sent an AI‑generated image of someone supremely bored by policy manuals.

At Alliant, we’ve also been diving deeply into the latest AI technology. We expect that over the next few years, AI will transform every aspect of the benefits universe, from insurance underwriting to employee communication. Accordingly, we’re monitoring these initiatives closely so we can advise employers on how to take advantage of the advances in AI technology while avoiding its pitfalls. We’re also closely watching how AI is advancing healthcare by improving diagnosis, treatment, and drug discovery.

For us, it’s not enough simply to observe others use AI. We are building our own AI-based systems, which will enable us to deliver better advice and gain a deeper understanding of AI’s capabilities. Much of this work is spearheaded by our employee benefits Analytics group, which has several projects that use AI to help clients optimize their benefits offerings.

“AI is not just a buzzword for us at Alliant. It’s a tool that will enable us to sift through vast amounts of data and extract insights that would otherwise be missed or require significant effort to uncover,” explains Doug Levit, Alliant’s executive vice president and director of Analytics. “We’re not just looking for obvious answers, we’re using AI to uncover profound implications that can transform our understanding of benefits and healthcare.”

The potential for AI in Analytics

The real potential of AI for analytics lies in machine learning (ML) technology. Unlike the large language models (LLMs) used in conversational systems like Chat GPT, ML systems are specialists that can be trained to deliver sophisticated insights from a detailed controlled data set, opening up exciting possibilities for the benefits industry.

As an example, here are two applications for machine learning technology:

Detecting Employee Health Trends

An AI model that spots emerging patterns in our extensive data about members’ health conditions and care will enable employers to better identify and support their employees’ health needs and get a more accurate picture of future benefit costs. For example, today’s models can identify members who have been diagnosed with diabetes. ML models are working toward identifying members with a higher probability of developing diabetes in the future.

Recommending Optimal Benefits Programs

Alliant is also exploring ML models that will be trained on the thousands of benefit programs and features that are available to employers. When used with a model that predicts member health needs, this system could help identify the benefits package best suited to each company’s employee base. We will even be able to help employers target certain solutions to members in specific regions, demographic groups, or other subtle factors that the AI can identify.

Vision for the Future

Even as we build these applications, we’re monitoring the rapid evolution of AI technology. In the coming years, we expect AI to make more accurate inferences from complex data and improve nearly every aspect of employee benefits, especially in these three areas:

• Personalized member solutions. As they enroll in and use their benefits, members will get recommendations for the best options based on their objectives, conditions, providers, and family configurations.
• Enhanced predictive modeling. Employers will be able to count on more accurate estimates of future benefits costs as models learn to make more fine-grained predictions of the condition and prognosis of individual members.
• Improved healthcare efficiency. Employers and members will benefit as AI is woven into the healthcare system. Conditions will be diagnosed earlier, and treatments will be optimized. AI will also be able to improve the communication between providers, health systems, and pharmacies by highlighting the most significant information.

Where AI differs from earlier approaches to data analytics is its capacity to learn from experience. Over time, these models will incorporate feedback signals that will improve their accuracy and ultimately enable them to discern more subtle patterns.

The work we’re doing now will help us take advantage of the deeper insights that AI will generate, empowering our clients to offer benefits programs that are smarter, more efficient, and more personal.

As Levit explains, “We’ll be able to spot risks that are bubbling under the surface so our clients can get ahead of the situation and put in a solution before the problem develops.”