Anthony Onesto (spoken monologue) argues that employee benefits are a critical but often invisible HR function—like plumbing or electricity—and that AI can substantially improve benefits through trend analysis, personalization, automation of administration, and data-driven plan design. He notes Suzy’s self-insured data (aggregate drug and service usage) as enabling analytics and gives an example linking the drug Wegovy to decreased diabetes costs as a use-case for evaluating plan design and cost-effectiveness.
Chapters
Anthony. It's Anthony Ernesto. And this is my AI and HR drop for this week. This week I'm talking about employee benefits. When I think of benefits within HR functions, I often refer to it as plumbing and electricity. The reason why is a lot of times you don't know they're there until something goes wrong. If you think about payroll or benefits, it really exists and it's just there. And then when you need it and it's not working, it's, it's terribly disruptive. Think about walking into your apartment this evening, turning on the lights and the lights don't go on, right? Otherwise if they go on, you don't even recognize that it's a thing. But when they don't go on, it's terribly disruptive. And so when I think of benefits, I think of them as plumbing and electricity and critical components often forgotten about within the HR functions. And I think there's an incredible opportunity for AI to impact this industry specifically because a lot of benefits, even if you're community rated, has a lot of opportunity to look at trends. What people are using, what benefits are consistently being used, what's not being used. And I think AI is going to be fundamental in helping organizations. Now at Suzy, we went self insured, so we have access to all the data, we have access to drug uses, what services are being used, obviously at a, not at an individual level, but at an aggregate level. And that will allow us to look at these things. Now if you throw all this data obviously in an AI solution that's either run locally, that's private to your organization, or something that you contractually buy, like a ChatGPT enterprise where you're not putting sensitive data like this into the Internet or even into the model, you could do some really amazing things. AI can look at this and understand trends. I'll give you an example. So what Gobi is super expensive drug to allow in your insurance benefits. But we see that there is a direct correlation of WeGovy and a decrease in any sort of diabetes. Now it's too early to tell. I mean, that's ultimately what the drug itself perpetuates. But if you can make a connection like that, where it's like, hey, our diabetes or preventative diabetes care is so expensive, it's not even close to what wegovy would cost you. So I think that's a perfect example of data analysis. I think you could personalize the benefit experience so that way different employees at different stages of their life find benefits important in different ways. You can automate a lot of the benefits administration. Think about all the elements, all these disconnected and disparate systems, and then you can really do data driven plan design, as I mentioned before. So start thinking about how AI could be helpful, particularly with your benefits administration. That's my quick drop. We'll see you next time.
