While artificial intelligence isn’t a novel concept in benefits administration, all sides of the industry are still considering best uses and areas for improvement.
1Password’s benefits provider incorporates AI for data analytics, as well as typical service requests and personalization, says Katya Laviolette, the software company’s chief people officer. For employees looking for different resources, there are planned texts and frequently asked questions to help streamline these searches, she notes, while its claims process has become much faster with a very minimal wait time for reimbursement compared to a few years ago.
Sobeys Inc.’s providers use AI to send instant text responses to employees to help them understand their benefits and what’s available to them, says Basil Rowe, the company’s vice-president of total rewards. “Our plan members are looking for quick answers to their questions, so being able to search on their phone and get the answer through the app or a call centre [is more efficient]. It also cuts down on the call volume and the time it takes to find relevant information.”
Read: Employers seeking simpler benefits solutions amid increasing HR demands: survey
Additionally, the grocery retailer is using AI for benefits analytics and identifying how to evolve its plan. “Benefits, in particular, is a very complex space, with everything from health to vision to dental coverage,” he says. “[With AI, we can] look at member usage in the plan and how our plan should evolve and it can inform future plan design considerations. That’s where we’re seeing the real value of analytics coming from AI.”
These forms of AI have been available at Sobeys for the last two years and their evolution is apparent in the company’s annual review with its benefits provider, involving more in-depth analysis, anticipatory questions and clearly identified trends.
Insurer applications
Beneva Inc. believes AI should be used to find efficiency in every task and deliver a better experience for plan sponsors and members, according to Sunil Hirjee, its vice-president of group sales and partner experience.
It’s well integrated into Beneva’s strategic plan, he adds, so much so that the insurer created a vice-president role that looks at how AI can be leveraged to augment the partner experience.
“We see AI as an opportunity for us to improve everything we do, [but] it won’t replace people. It’s just being used to make things easier, faster, more accurate and free up time for us to focus more on the human aspect of what we do.”
Read: 2024 Group Benefits Providers Report: Leveraging AI to streamline benefits
In addition, Beneva is using AI through its call centres for transcription services to ensure there’s an accurate record of phone call details, as well as call routing so people can get to the right resources immediately. It’s also looking at how it can evolve so members have the ability to engage with a virtual agent for simple, frequent questions.
At Medavie Blue Cross, AI is integrated across benefits administration to reduce friction, improve accuracy and give plan sponsors clearer insights into plan performance and member needs, says David Adams, the company’s senior vice-president of insurance business. “This includes intelligent document processing for faster, more consistent claims handling, agentic systems that automate multi-step operational workflows and AI enhanced reporting that supports plan sponsors with real time, actionable data.”
Medavie also applies AI to fraud detection and cybersecurity to help improve plan sustainability, he notes, adding its AI capabilities have steadily expanded over the past several years, with more than 80 active AI use cases underway across the organization.
While Beneva started incorporating AI into plan member services in the last 18 months, its use in fraud detection has been around slightly longer, says Hirjee. “One of the major strengths of AI innovations is the ability to identify trends, patterns and anomalies. It’s already integrated into fraud prevention and abuse strategies at Beneva to ensure we’re doing our best to keep the premium sustainable for plan sponsors.”
Read: AI tools helping insurers manage plan costs by weeding out instances of benefits fraud
Medavie’s AI applications are meaningfully reducing manual work and simplifying administration for plan sponsors as well as its internal team, says Adams. “By automating routine tasks and improving data accuracy, we’ve increased processing speed while allowing employees to focus on complex situations where human reassurance matters most. Our digital advancements also improve transparency and navigation, giving plan sponsors and brokers clearer insights.”
It’s too soon to measure whether Beneva’s AI features are having a significant impact for members, says Hirjee, but the organization has run some pilot projects internally that clearly identified time savings. A recent pilot for a specific group of employees found a measured gain of about two hours per week.
User feedback
Sobeys’ employees seem open to the use of AI for their benefits, as long as it’s making their lives easier, says Rowe.
Key takeaways
• Benefits providers are harnessing AI to make the plan member experience more efficient and to give employees the freedom to focus on more important tasks, not replace jobs.
• Plan members appreciate the speed of claims and finding resources thanks to AI, but want more personalization and are still concerned about privacy.
• AI will be further integrated into benefits administration moving forward, but the focus will be on security and how to maintain the human aspect to protect people’s health and well-being.
“They want to interact with [their benefits] the same way they interact with the rest of the world. They’re looking things up online and getting insightful answers, so they expect the same thing [with] their benefits. Employees also expect we’re doing the right things to protect their information and we work with our providers to [ensure it’s protected].”
However, at 1Password, Laviolette says there are two schools of thought when it comes to employees’ perspectives on the use of AI and benefits. Some plan members appreciate the efficiency gains, fast reimbursements and ease of access to information, while others are concerned about the privacy risks.
Read: Editorial: Looking at the good, bad and ugly of artificial intelligence
“The other side is the nervousness of security and making sure we’re protecting personal information. We’re a security company and it’s very important that we protect people’s identities, so when we’re working with our providers, we want to ensure they meet our standards as well.”
1Password receives employee feedback on benefits administration during its annual reviews and constantly shares feedback through its broker to keep providers informed on what works and what doesn’t, she adds.
Plan sponsor feedback is critical, especially from a security and privacy perspective, says Dimitri Poliak, a principal at Normandin Beaudry. “I think the biggest fear around agentic solutions is the sandbox in which they operate. Because there’s the ability to just let it do whatever it wants as long as it can get from A to B. But if you’re not careful, it will do things behind the scenes that are less desirable.”
There are currently two opposing forces in the market, says Francis Boulianne, a principal at Normandin Beaudry. “We’re seeing a lot of clients getting stricter about privacy and security, but on the other side, we have all of these AI tools improving access and wait times. Personal information related to claims data is protected by regulation, [but the AI function might not recognize this]. The importance of client feedback is advocating for a strong framework for all benefits users.”
The biggest issue the industry is trying to address is the fundamentals, says Poliak. As exciting as agentic AI may be, the focus should be on getting the basics right, he adds, making sure data integrity is maintained and administrative processes are executed correctly.
Read: AI supporting, but not replacing, pension, benefits teams: experts
“The issue with any introduction of a new technology . . . is we need to be mindful of understanding the process behind the scenes. It’s one thing to train the technology to do human functions, . . . but we need to understand the process behind it to ensure compliance, data integrity, but also to ensure security. If we move away from having experts in the field being able to explain or understand the process, this is a concern that will grow over time.”
Another issue is the fragmented structure on the provider side, adds Boulianne. “Clients [often] have multiple providers involved, so they need to create [a structure] that’s consolidated for employees and think of how to bring this complex ecosystem together. Right now, I think the siloed approach of developing those systems is the main blocker.”
For the user experience, plan members want to talk to an agent and receive appropriate guidance that leads to action, but it’s much harder to do than simply flipping a switch, says Poliak, because it involves a variety of systems in a variety of benefits.
“Even within providers, they have different teams responsible for different functions, so it’s not just a surface layer connectivity. It has to permeate all the way through the organization, [looking at], ‘What are the effects if a transaction is placed?’ and ‘How is the administrative rule going to be brought forward to the agent level?’”
Plan member perspective
For 1Password employees in Canada, most of the AI-enabled features are built into the benefits provider’s mobile app and online portal, says Kiya Finch, manager of people experience and growth.
It can also be used to help locate nearby providers, such as physiotherapists offering direct billing, which makes the process more seamless. “For straightforward tasks, the tools work very well,” she says, speaking from the perspective of a plan member. “The claims process is fast and being able to submit a receipt in seconds removes friction. Reimbursements are typically processed quickly and that convenience increases the likelihood of employees actually using their benefits.”
Read: 2022 Tech Insights: How digital administration can improve benefits plan access for employees
Reminders about eligible services are particularly useful, adds Finch, especially for benefits employees may not realize are available. “There’s a virtual assistant that answers common questions about coverage, eligibility and claims. The claims process is largely automated [and the system] often processes reimbursement without human review.”
1Password’s provider has a virtual assistant that’s effective for simple questions, but when they become more situational or nuanced, live support tends to be more helpful, she notes, adding the biggest opportunity for improvement is around deeper personalization.
“AI has the potential to better reflect life stage, previous claims patterns and plan details to provide more proactive guidance — almost like a benefits concierge. For example, after submitting a physiotherapy claim, it would be helpful to automatically see how many visits remain or receive suggestions for related covered services.”
Escalation is another area for improvement, says Finch, noting when AI can’t resolve an issue, the transition to a human representative should be seamless without requiring information to be repeated. “AI has already made benefits administration faster and more accessible. The next step is making the experience more intuitive, contextual and human-centred.”
What does the future hold?
Beneva is continuing to look at ways to use AI moving forward, but no matter what happens, it isn’t going to replace the human aspect in how insurers operate, says Hirjee.
“The reality is AI is going to profoundly transform the way companies operate in general and I don’t think there’s a single business sector that’s going to be exempt from that. But [it has to be approached with caution because] it’s unlike other technologies we’ve implemented in the past.”
Read: Plan sponsors focusing on costs, member feedback when redesigning benefits programs: expert
The insurer is encouraging its employees to use some of the available solutions and truly think about how they want to work with AI, he adds, noting there’s going to be a strong focus on improved utilization and innovation over the next six months.
Looking to the future, Medavie is focused on expanding responsible, human-centred AI that enhances navigation and provides proactive insights for plan sponsors and members, says Adams. “This includes deeper predictive analytics, more integrated support across digital and inperson care and continued improvements to conversational AI to guide members through complex benefits questions. Our goal is not to replace human expertise, but to give our people better tools so they can support members with greater clarity and empathy.”
A lot of employee-facing systems still need to be updated to allow for consolidation, notes Boulianne. “In the back end, there are still a lot of older, fragmented systems that aren’t necessarily supportive of new technology. [To make improvements, providers and] plan sponsors have to listen and understand what plan members need [and make the necessary updates].”
For Sobeys, the real opportunity to harness AI is by partnering with its benefits provider and vendors to ensure the plan member experience is seamless, says Rowe. “The reality is there may be handoffs in the background, but we want to make the process as smooth as possible for employees. So we’re always working on it from a member perspective — how to provide that one-window approach and then we can triage and stream the information to the appropriate vendor to action it.”
In the machine learning space, AI has been around for a long time, notes Laviolette, but it’s constantly transforming. “I always say, if you’re not embracing it or being curious about it, it’s going to pass you by. It’s transforming businesses, it’s transforming jobs, but there’s also an aspect of humanness in benefits that’s really important, so at what point do you ensure there’s some type of human interaction?”
Sadie Janes is an associate editor at Benefits Canada and the Canadian Investment Review.