A Conversation with Jason Arvanites, Camber’s New VP of Operations

Specialty care revenue is a uniquely difficult problem. There’s a lot of complexity in the billing process, a huge amount of administrative work, and too often, the financial outcome doesn’t reflect the quality of care being delivered.

The last part really stuck with me. You have clinicians focused on doing great work for patients, while a lot of the infrastructure around getting paid is still incredibly manual and fragmented. I started to think there was an opportunity to use technology to take some of that burden off their plates and make the economics of running a practice work better. It felt like a meaningful problem to take on next.

Honestly, I didn’t fully realize how critical it is to minimize the number of people who need to touch a claim. I came in thinking a lot about speed and throughput, but the more I looked at the process, the more I saw how much both depend on reducing handoffs.

Every time a claim gets picked up, someone has to understand the context, figure out what’s already happened, and decide what to do next. Each handoff adds complexity and another opportunity for something to get missed, misunderstood, or done twice. Even if each person only spends a few minutes on it, that adds up quickly.

No one necessarily dropped the ball. The claim just had too many stops along the way. So I keep coming back to the question: what’s the fewest number of people who need to touch a claim to get it right? I think the answer is in how you design the process.

These are hard businesses to build and run because you have to get two very different things right at the same time: building good software and running a good operation. You can’t really separate the two.

One thing that’s especially interesting at Camber is that the people and teams using the product every day aren’t necessarily the people buying it. Our Insurance Operations team is in the product every day, making decisions, feeding what they learn back into our data model, and moving claims forward. The clinic sees very little of that work. They see the outcome: whether claims get paid, how quickly they get paid, and how much work they have to put in along the way.

So the operational metrics become a really important measure of product quality. Engineering needs to understand those metrics and the operation behind them, otherwise you can end up building something that works really well internally but doesn’t materially change the customer experience.

They underestimate how much complexity they’re carrying. It’s relatively easy to find a workaround for a messy process. The hard part is stepping back and figuring out why it’s messy in the first place.

I see a lot of teams get very good at managing that complexity. They know the exceptions, the workarounds, and who to call when something goes wrong. And that can work for a long time, until volume grows, something changes, or the person who knows all the unwritten rules moves on. Then you realize how much of the process was being held together by people rather than the system.

I’d say there are two.

The first is figuring out how quickly you can take work that’s being done manually today, or where someone has to check every step, and move it to a place where that person is overseeing the work instead of sitting in the middle of it. That’s true for us and our customers. There’s a lot of time being spent on work that doesn’t really need a human involved every time.

The second follows pretty naturally from that. Once you automate the routine work, what’s left is the harder stuff. And as you keep automating, that work gets more complex. So the challenge becomes how quickly you can hire and train people who can handle those harder claims both accurately and efficiently. Automation doesn’t make that constraint go away, and if anything, it makes it more important.

I’m only a few weeks in, so I’m still forming a view on this, but my early observation is that not all automation creates the same kind of impact.

There are plenty of tasks that are good candidates for automation because they’re repetitive and relatively easy to measure. Eligibility, claim scrubbing, payer status checks, and some straightforward denial work fall into that category.

But if the next step is still “send it to a person to check,” you haven’t necessarily created much operational capacity. You’ve really just made one part of the process faster while keeping the same amount of work in the system.

The other unlock I see is in the part of the product our customers don’t really see. AI can look across thousands of claims and surface patterns, like which steps are driving rework or where a rule or SOP may have gone stale. That gives our teams visibility into problems that would be really hard to catch looking at claims one at a time, and sometimes, fixing the underlying process is more valuable than just automating around it.

What makes revenue cycle particularly challenging is that the process is constantly changing. New payers, policies, states, and clinic-specific requirements create new exceptions all the time. So I think the real opportunity with AI is less about automating a fixed list of tasks and more about building a system that can handle more of that variability without continuously adding people to manage it.

Compliance, and more specifically how much of the burden landed on the customer. We did what we needed to do to keep our customers accredited, and on that measure we delivered. But getting there asked a lot of them. A documentation request would land, we'd go back to them for what we needed, something would surface, and we'd work through it together. From their side that meant a steady stream of asks, not much visibility into where things stood, and a fair amount of anxiety about whether it was all going to come together.

The outcome and the experience are two different things, and we only really finished the first one. You can hand someone the outcome they paid for and still put them through a frustrating time of getting there.

It's also part of what I've appreciated about Camber so far. Compliance is something we're thinking about as we build the product and the operation. Given how important it is to healthcare billing, that feels like the right way to build.

Predictability. I’d want a customer to feel like their revenue cycle is becoming something they can rely on, regardless of their geography, clinic footprint, or payer mix.

From an operations perspective, success is also about reducing how much attention the revenue cycle requires. If customers are spending fewer hours reviewing claims, answering questions, or figuring out why something happened, that’s a meaningful improvement.

The ideal outcome is that it becomes less visible. Cash arrives when it should, exceptions are handled without constantly pulling the customer in, and the revenue cycle becomes infrastructure they can trust to do the right thing time and time again, rather than something they have to actively manage.

Curiosity, first and foremost. The best operators I’ve worked with want to understand the whole system. They don’t just ask what happened. They want to know why it happened, what caused it, and what we can learn from it. That curiosity tends to lead to better decisions and better solutions.

I also look for hunger. Building something great takes a lot of persistence, and I want people who have a real desire to do excellent work. The source of that motivation can be different from person to person. What matters is that it’s there.

And integrity is non-negotiable. Operations is full of moments where the easy answer and the right answer aren’t the same. I want people I can trust to be honest about what’s happening, own their mistakes, and make the right call when nobody is watching.

Teaching is still the hardest job I’ve had. There’s nowhere to hide, and kids notice everything. Three years in the classroom taught me the importance of deep preparation and taking ownership of the outcome, but also that you can’t solve a problem without understanding it from the other person’s perspective. You can have a great plan, but if you don’t understand what the person in front of you needs, it’s probably not going to work.

At Bain, I learned how to apply some of those same lessons in a very different environment. I learned how to bring structure to ambiguous problems, work through them with a team, and get comfortable with a lot of feedback. In both settings, you’re ultimately trying to understand what’s really happening, figure out what needs to change, and take responsibility for getting there.

Jason is still early in his Camber journey, but he’s spent much of his career thinking about how operations need to evolve as a business grows. One thing seems to have held true across each of his stops: the hardest operational problems are rarely solved by technology alone.

They’re solved by understanding how the work actually gets done, where the real constraints are, and what needs to change for the organization to operate better as it grows. Jason has spent his career working on exactly these kinds of problems, and we’re so excited to have him on board as he takes on this challenge at Camber.