Revenue Cycle Insider

Learn how AI can scan your denials to identify high-value appeals.

Adding artificial intelligence (AI) to your practice can be an empowering experience that streamlines your workflows and allows humans to handle more complex tasks. However, could the technology be causing headaches that require more attention than originally thought?

During the session “Beyond the Algorithm: Redefining Human Power in Coding & Revenue Cycle,” at AAPC’s HEALTHCON 2026, Shannon C. Cameron, MBA, MHIIM, CPC, chief operating officer for AFS of Harvard Medical Faculty Physicians, examined where you’ll see the costs of AI in the revenue cycle.

Keep reading to learn where AI fits best into the revenue cycle and its costs to your organization.

Understand How AI Has Changed the Revenue Cycle

AI has altered the revenue cycle in positive and negative ways. For one, there’s no doubt that AI accelerates throughput. The technology triages work, identifies patterns, and flags anomalies quickly. AI analyzes simple cases and suggests codes that line up with the documentation, which allows coders and billers to move swiftly through workflows.

The technology can also scale decision making up or down depending on your organization’s needs. However, that includes both good and bad decisions. “If your policies are misaligned with payer interpretation, AI is going to enforce that misalignment, and it’s going to be consistent. If you’re in a big organization, there could be a lot of claims going out the door with the same problem,” Cameron said.

If your organization is submitting claims that are incorrect due to errors in the AI, you have a big problem. For instance, your dashboards may show high rates of clean claims while medical necessity issues are brewing under the surface. Without governance, clinical judgment, or financial oversight, the AI could be multiplying problems that will result in your practice issuing refunds or paying penalties.

“If implemented correctly, AI can prevent loss, but if implemented poorly, the technology institutionalizes your risk. We’ve got to find that balance,” Cameron explained.

Find out Where AI Creates Real Economic Value

According to Cameron, the true economic value of AI is experienced before claims are submitted to the payers. Upstream use of AI helps solidify your claim before it heads out the door. Below are the various ways that AI streamlines your claim submissions:

• Predicts denial risk: One of the biggest issues in the revenue cycle is working claims incorrectly, explained Cameron. For example, a lot of organizations appeal denials no matter what, even though some denials don’t need to be appealed. Cameron said that the same methodology can apply to write-offs. Certain claim adjustment reason codes (CARCs), remittance advice remark codes (RARCs), and sequestrations don’t require humans to perform manual work. “We used AI to write sequestrations off, so I could use our accounts receivable [AR] team and their human insight on more complicated types of denials,” Cameron said.
• Validates medical necessity before billing: When providers are assigned to perform a service, rather than scheduling a procedure, you might receive a case where the medical policy wasn’t met. An AI system can identify the applicable patient diagnoses against diagnosis lists from payers to ensure the claims are accurate. “We were able to use payer-specific information and feed it back to all the points of reference and the patient access, and give that information back upon scheduling,” Cameron explained. She added that if you have the right AI system in place, the technology can prevent billing issues from arising even before the appointment.
• Prioritize appeals by ROI: As you examine your appeals, you might wonder if the cost to collect reimbursement is worth the effort. Your appeals list might have cases that shouldn’t be appealed. This is where AI can help identify the cases that are worth the money. “If you’re looking at it from work smarter, not harder, that’s where AI can come in, and those things can be programmed for you,” Cameron said.
• Identify contract variances: Regardless of how you track reimbursement from payers, there are sure to be differences between the payers. Payers may use CARCs to deny cases, and if your AI is set up to automatically write claims off when they’re denied with applicable CARCs, then you could be losing out on revenue if you’re not reviewing denied cases that actually should be appealed. “If the payer is denying claims above the allowable or they’re denying 100 percent of claims, you should have a process in place. If you’re using AI, then those processes can be built into your system,” Cameron said.

Implementing AI as part of the start of your revenue cycle can help reduce headaches and minimize hurdles to reimbursement later on, but you need to be aware of what obstacles lie ahead.

Does AI Quietly Increase Cost?

As productive and efficient as AI can be in your organization, the technology can also quietly increase your operating costs. If your team deploys AI and automates several revenue cycle processes, but no one examines its performance over time, you could be setting your organization up for problems.

“We hear audit and we think of coding audits, but we should always be auditing the entire revenue cycle process, from write-offs to your payments, to any part, any piece or part of your revenue,” Cameron stated.

Plus, as mentioned above, AI could automatically appeal denials that might not need to be appealed. This increased appeal volume will increase your costs mostly from wasted time spent trying to recoup payment for claims that should be written off.

Don’t forget that if you have a compliance issue that gets baked into your AI system, that problem will compound when the system is scaled and then you could have a compliance issue at a much bigger level. Remember that AI amplifies whatever processes your organization already has, even the weak processes.

Simultaneously, you need to take steps to audit your vendors to avoid the risk of dependency on third parties. If you allow the vendors to have the responsibility of building a special workflow for your team or installing a feature into the system without oversight, you’re leaving the decision making up to them. You should audit how the system works and you need to test what vendors have done to the AI platform to ensure the system and any adjustments are compliant.

Remember That Adding AI is a Marathon, Not a Sprint

Deploying AI in the revenue cycle can be a huge improvement in your team’s throughput, but looking at the tool as a cost-cutting measure is shortsighted. “It’s a short-term mindset because what you gain in cost savings, you might lose in compliance exposure and revenue leakage,” Cameron added.

Mike Shaughnessy, BA, CPC, Production Editor, AAPC