Let the Experts Handle It

Let’s start with the proof of concept that ReSource Pro is working on with this large employee benefits distributor. Why did the client pursue this?

AMBER BRETHOUWER: The client is very techsavvy. And this particular process, which is focused on middle-market and large groups, is very manual and time-consuming. Ultimately, what they really want is to provide the same level of technological sophistication for this complex process, and they want to be able to provide the same turnaround time that they deliver in the small market, where this is almost a real-time process.

ZACH HANSEN: It’s employee time savings. This specific client wants to be able to win deals to drive revenue growth. They want to sell more by delivering faster turnaround and better outcomes, but they have this very tactical process, creating plan comparison sheets, that must be done and slows everything down. It quickly becomes a scalability and human time problem. Given ReSource Pro’s position, we see a lot of volume. This allows us to help clients orient and transform their organizations for the era of agentic process management. This means building teams that act as expert supervisors of agentic work. And so that’s exactly what it is here.

What have been some obstacles to implementing this AI initiative?

BRETHOUWER: The first obstacle is change management. First, there’s proving that this is an activity that can be transformed; second, proving that this is an activity that once you apply agentic AI you can still count on a reliable, quality standard.

The ReSource Pro value proposition was first instilling confidence in our domain expertise, that we manage this process for nearly 100 clients today, and that we have a lot of experience doing it. And then building that into a model where our expert with that insurance domain expertise is the one operating the AI agent, to ensure the same quality of output, just at a much faster turnaround time and at a much lower price point. For us it’s about always maintaining an expert in the loop.

Lastly, when you think about the employee benefit space and this particular process of spreadsheeting plan comparisons, there’s very little standardization in the market. Sometimes there’s a view of you just dump it in AI and some magical thing appears on the back end. But there’s really a lot of learning and prompting that has to be done in this particular model so that it can overcome all the variations.

What advantages come with working with a third-party expert like ReSource Pro rather than businesses building their own AI solution?

HANSEN: It’s easy to get over your skis. What this client is really good at is distributing insurance. They’ve said to us, “Hey, you handle the operations.” It’s a much faster path to adopting AI because we have established trust as process management experts, specifically within insurance. We’ve made investments at scale to help our customers accelerate where and how they integrate AI into their operations. So instead of them diverting their resources to solve various acute problems, which takes a lot of time and testing and pain, they can rely on us and just focus on their core work.

On that same point, because we spent a lot of time building this agentic framework, we’ve already addressed security, privacy and governance issues.

BRETHOUWER: It’s kind of like buying a cake at a bakery versus buying the ingredients and trying to bake the cake by yourself at home. It seems to be resonating not just with this client but with others; they’re less concerned about how the cake gets baked and just want the cake.

What qualities should a company prioritize when choosing a partner to help with AI solutions?

HANSEN: Three things. One: insurance expertise. This is foundational and, in my opinion, non-negotiable. We can have another full interview on the many reasons why. Two: change management. This is a people business and to win, companies need technology and people to work well together. And three: measurable outcomes. Does the partner have a framework and a system for measuring inputs and outputs and generating the governance layer that you need? There are other criteria, but I would start with these.