Reimbursements in healthcare have been at the top of leader’s minds for decades, creating a burdensome process for providers, patients, and manufacturers.
Physicians are more hesitant to use certain devices, or perform certain procedures, if they don’t believe they will be reimbursed at a fair rate.
Reimbursement debates often cause heated arguments between providers, insurance plans, and players in the same, much like recent
drama
surrounding CMS coverage for TAVR.
To overcome these decades old healthcare challenges, companies are trying to figure out how to better use the AI tools at their disposal to make payment systems easier, faster, and less cumbersome for patients, providers, and devicemakers.
MD+DI
connected with Roshan Patel, founder and CEO of healthcare technology firm Arrow, to discuss AI’s role in medical billing, predictive analytics for cash flow operations, and more.
Arrow, formerly known as Walnut, emerged from stealth in 2024 following its founding in 2020, offering payments solutions featuring advanced error detection, predictive analytics, denial management, and comprehensive visibility into revenue cycles to streamline claim processes, boost claim accuracy, and drive smarter decision-making and operational efficiency.
The company secured a $110 million Series A funding round in 2022, and is backed by major players in the space including Google, Plaid, Mercury, and Ramp.
How is AI transforming medical billing and payments specifically for medical device companies and their revenue cycles?
Patel:
AI is impacting every industry. Billing and revenue cycle is a knowledge and information type of work, and so we are starting to see AI take a lot of the tedious and repetitive manual processes in getting an insurance claim paid, which to go from a patient visit to money in the bank there is a large arduous process with tons of friction, and that is usually by design by insurance companies. They don't make it very easy to get paid as a healthcare provider. To even just ask the status of a claim, sometimes you can go and check the portal if they have one, but many times you have to call. Half the time, revenue cycle teams are just calling and that's a huge time suck in their day, just being stuck on hold for 30 minutes. Something like that is very easy for AI to take over, so billing teams can focus on what actually requires their judgment and knowledge, and not just be bogged down with so much easy, repetitive manual work.
How can predictive analytics help companies better forecast payment timelines and cash flow?
Patel:
There is a ton of data healthcare providers are sitting on that they are not utilizing because they are already so busy and underwater. If you can put payment and revenue cycle data into a business intelligence tool, you can start to see a lot of trends and forecast things out. EMRs aren't really built for financial forecasting, they're built for clinical work. A lot of people are trying to find those tools to project out, so they can project out, ‘when I submit this claim to Aetna, what is the probability I will get paid and how long will it take to happen?’ And then you can start to build your business around that.
What payment friction points do you see most commonly?
Patel:
The most common one is denied claims. There is a statistic that roughly 20% of all claims get denied, and a lot of times that is due to errors in billing early in the process, like wrong patient insurance, coding errors, things like that. A huge chunk of the time it's an error on the insurance company's side because they have denied something incorrectly. Once a claim is denied, it kicks off a huge process on the billing team side, since they have to figure out why it was denied, was it a mistake, and what needs to be done to fix the claim. Then you get a back and forth with the insurance company, so that's where we are seeing most of the time suck, but also the most high return on their time, because they are getting an unpaid claim turned into a paid claim.
What trends are you seeing in patient financial responsibility, and how is this impacting the industry?
Patel:
Patients want more transparency, especially into what things will cost and assurance of that cost upfront before getting care. That is hard to provide that information, because so many things downstream affect how much the patient will know. If a claim is denied, we cannot predict that with 100% certainty. You can have a good chance of knowing, but you’ll never know what an insurance company will do. Also, the core treatment for a patient can change, making it hard to know what services will be rendered, making it hard to estimate costs upfront when you don’t know what will happen in a specific treatment.
What payment and billing trends should executives be watching in 2026?
Patel:
Everyone is evaluating AI and how to best use it in their workflows. There is AI for almost everything in payment and billing- every slice of the revenue cycle. I think it’s about trying to find one where, right now, everyone is promising the world, but how do you know what will work? I think that is the biggest challenge for executives. People are now simply asking their peers what systems they are using, and what is working. That is how things are getting adopted nowadays.