While there’s multiple opportunities for the use of artificial intelligence in the employee benefits space — including making the claims process more efficient and personalizing the plan member experience — one of the biggest advantages of the technology is the ability to streamline administrative tasks, allowing employees to focus on more value-added work in addition to their own growth and development, says Mara Notarfonzo, vice-president of total rewards at CAA Club Group.

She recognizes insurers are already implementing this efficiency through claims submissions and various online tools. “The claims process has definitely become more automated. Turnaround time for processing claims and getting funds reimbursed is much faster than it was in the past. I think the main thing carriers are starting to do well — from a user experience — is using AI to simplify things.”

While AI itself has been around for some time, the technology is currently having a moment in the employee benefits sector due to its ease of use, says Tom Milne, independent consultant for Seven Studios.

Read: ChatGPT a game-changer for benefits communication, but risks remain with full automation: expert

“Things like ChatGPT are improving member communication and can relieve some [administrative duties] around answering basic questions. The other major opportunity related to benefits is crunching data. Benefits plans rely on data, whether it’s related to communication and . . . maintaining email lists when you get bounce-backs, for example. Administrators can focus on more of the problem solving or strategic tasks, while AI helps with those mundane administrative tasks.”

With proper legal approvals, insurers could improve plan member education by using AI to develop a specific questionnaire — around factors such as age, income and health history — that could easily show the benefits that were chosen by plan members in similar situations, he says. Not only would this help new plan members, but it could support those whose personal circumstances have changed or those who made inadequate selections based on limited knowledge of the plan.

“[Insurers] have been focused on getting the most relevant information to members in the least amount of clicks. They build the menu and content based on past experience of human navigation, but continually monitor analytics and adapt to ensure the user is finding the information they need. With improved AI search engines, they can put the user in the driver’s seat.”

Insurer implementation

AI can certainly provide a competitive edge for insurers if they harness it properly, notes Milne, adding insurers are often using it to mine their data for health-care trends that can aid companies in financial planning or to make plan changes that help control costs.

Such tools can help enhance health outcomes and guide members to make more informed decisions, says Ashesh Desai, head of group benefits at Manulife Financial Corp. “We’re looking at how to make things easier for our call centre agents and people [who] engage with our members directly. AI has been able to help us . . . pull together all the data and serve it up to that call centre agent in an easy-to-access fashion.”

Read: Humi using ChatGPT to remove administrative burden from HR, recruitment processes

There’s also a significant opportunity to leverage AI to better understand what plan members need. “AI has allowed us to extract what we call ‘the voice of a customer’ — extracting common themes and sentiment for us to [identify] pain points we should address or better understand . . .[members’ feedback].”

Using this information, Manulife is able to develop programs, enhance member journeys and remove some of the friction that may exist to better serve its clients. “If you think about the wealth of data that organizations like Manulife have, how do you pull all of that together to be able to actually enhance the health outcome?” asks Desai. “Many disability claims have to do with mental health, so we want to . . . intervene sooner, provide access to care and engage with members in a more efficient way. We’re able to pluck out some of those journeys based on health, drug and disability data and recognize you can decrease people taking certain medications if [they] engage with a psychotherapist sooner.”

Manulife is in the midst of launching a new health platform using AI technology, currently piloting it among staff and launching it to all plan sponsors by mid-year. Through the platform, the insurer aims to engage with members, recognize the medication and type of therapy they’re using or looking for and make alternative suggestions, notes Desai.

“AI is a tool that our health platform is going to leverage in order to communicate with people. The platform basically brings benefits and health care together so members can find the care they need when they need it. It will be taking the online experience we already provide and enhancing it, then leveraging AI to serve up the right information at the right time.”

Read: 2023 Future of Work Summit: What do employers need to know about the use of AI in the workplace?

Key Takeaways

• AI can help streamline administrative tasks in the benefits space and simplify the claims process, but it can also be used to detect benefits fraud and create cost savings for plan members and plan sponsors.

• Insurers must be careful that generative AI data is used solely for its intended purpose, protects members’ confidentiality and doesn’t introduce bias or discrimination.

• AI is constantly evolving, so insurers are always adapting and learning how to best harness the technology responsibly.

Indeed, Notarfonzo says insurers seem to be on the cusp of helping plan members with benefits selection through AI. “[Many] employees don’t fully understand how health benefits and insurance work. Organizations have different coverage and, depending on the plan, you may have a co-share of premiums. With those types of plans, employees tend to make decisions based on the cost. So, I think there’s an opportunity to help people make informed decisions on the coverage they need.”

Beneva Inc. is currently building on its AI foundations to deliver value even faster, using the technology to improve member communication and collect data that can assist with chronic diseases and preventative care, says Étienne Plante-Dubé, senior director of advanced analytics and AI. “Prevention is a very important aspect when it comes to chronic diseases. We offer health promotion tools that can be personalized thanks to [AI]. For example, it is possible to predict a person’s return to work based on different types of support, and thus optimize support to the individual’s needs. It is also possible to offer personalized preventive advice and prevent the risk of disability relapses.”

AI can also help to better predict the costs of drugs that will be consumed in the future. “We can use AI to anticipate changes in medication intake and better predict future drug costs. AI could help us guide pharmacists towards less intuitive alternatives based on data analysis. This information could also better equip our members in our prevention and health promotion program by linking to more information.”

Insurers are also starting to leverage AI from a fraud and risk perspective, says Plante-Dubé, noting it can improve fraud detection and group insurance underwriting, which will have a positive effect on plan costs.

Possible concerns

Despite the opportunities for cost savings and efficiency, there are concerns around the use of AI, including digital security, says Notarfonzo.

Read: Survey finds Canadian employers looking at AI to enhance worker productivity, but concerns remain

“Digital security and personal trust go hand in hand. AI is evolving daily, so we need to make sure that we understand what the capabilities are while still safeguarding information. We also can’t rely on it too much because there’s got to be a balance between system-generated things and personal connections.”

Indeed, recent research from Beneva found there’s a great deal of public concern around the transparency of the algorithms used and the protection of personal information. The insurer collaborates with academic partners to better explain the behaviours of AI models and validate the meaning in summaries obtained through generative AI algorithms.

There could be several ethical issues related to the use of AI models like algorithmic discrimination and data confidentiality, adds Plante-Dubé. “That’s why we’ve put in place rigorous processes for assessing issues at every stage of our AI projects. We’ve also put in place a code of conduct for the responsible use of data, based on consensus principles for the responsible development of AI.”

In order to avoid security and trust issues, Milne says insurers need to constantly confirm that the AI system is learning. “I think that’s the biggest risk of saying, ‘We’re using AI and we want to be new and hip, and we’re at the forefront of technology,’ but when it doesn’t work, it’s worse for your members. Then you can have trust issues and complications with comunication. If you’re going to do it, you have to do it right.”

Manulife is also committed to the ethical and responsible use of AI throughout their organization, notes Desai, with multiple levels of governance and oversight to ensure that data is being used for its intended purpose.

Read: AI tools helping insurers manage plan costs by weeding out instances of benefits fraud

“I think especially in our space, we want to ensure we test . . . and we make sure the data is used for the appropriate intent. [There are] a few ways we’re looking at leveraging that data, but we also want to make sure nobody is disadvantaged by some of the insights AI would provide us.”

Future possibilities

Looking to the future of AI, Notarfonzo hopes to see insurers prioritize the personalization of benefits.

“Carriers have a wealth of knowledge that employers don’t have. From a wellness pespective, when we’re developing our holistic wellness programs, we’re developing them for the masses. We’re taking a look at information we receive at a high-level generic perspective. Wouldn’t it be great if carriers could leverage the data they have from drug claims to maybe predict where there’s opportunity to send personalized communications to people?”

Desai notes the new platform at Manulife could provide opportunities in terms of benefits literacy moving forward. “People know what benefits they have, but they may not have the details when they need it. How do you get ahead of that so people aren’t searching for that information or having a negative experience? By leveraging our data insights that are coming out of these AI-driven models, we can figure out what questions to ask our members.”

Beneva’s teams are continually working to better understand customers’ needs and improve its processes through AI, notes Plante-Dubé, adding the process for submitting a claim can still be improved. Members are required to enter a lot of information, so AI could be used to pre-fill several fields with automated verification and instant detection of potential fraud.

As long as humans stay in control of AI, it can be used to the industry’s advantage, says Milne. “As long as companies continue to analyze the risk and give direction to employees on how to use it, it can be beneficial. To protect personal health information or intellectual property, companies must continue looking from a legal perspective at what risks AI poses within their organizations, and what risk is associated with the information being given to employees.”

Sadie Janes is an associate editor at Benefits Canada and the Canadian Investment Review.

Download a PDF of the 2024 Group Benefits Providers Report.