GI Billing and RCM: Cut Claim Denials With AI Follow-Up

GI billing has a reputation among coders for a reason. A screening colonoscopy that turns diagnostic mid-procedure changes the modifier, the patient’s cost share, and sometimes the payer’s coverage rules – all after the scope is already out. Multiply that by anesthesia claims, pathology on removed polyps, and facility fees, and a single procedure can generate three or four claims that each have their own way of getting denied.

That complexity is manageable when someone works the claims on time. The problem in most gastroenterology practices is not that the billing team cannot handle the rules. It is that the volume of routine follow-up calls buries the day, and the claims that need a phone call to a payer or a patient sit untouched until a deadline closes.

Where GI revenue leaks between the scope and the deposit

Every step between the procedure and the payment is a place a claim can stall. Prior authorization that was approved for a screening but not the diagnostic conversion. A pathology charge that posts days later and never gets linked. A patient who owes a coinsurance amount they did not expect and lets three statements pile up before anyone calls.

Denials are a structural cost, not an occasional accident. Review of in-network claims in ACA marketplace plans found insurers denied a substantial share of submitted claims, and the overwhelming majority were never appealed (KFF). In a procedure-heavy specialty, each of those un-appealed denials represents a facility or professional fee that the practice earned and then abandoned because no one had the hours to fight for it.

The follow-up itself is not hard work. It is checking claim status, reading back a denial code, confirming a corrected claim was received, and calling patients about balances. It is repetitive phone-and-portal labor that any organized person can do, which is exactly why it loses every competition for a biller’s attention.

How an AI voice agent keeps the follow-up moving

Pretty Good AI builds voice agents that handle the administrative calls in your revenue cycle, integrated with athenahealth. The agent sits in front of your billing team as a logistics layer. It does not make coding decisions and it does not exercise any clinical judgment. It works the calls that keep claims and balances from aging out.

For payers, the agent runs outbound status checks on pended and denied claims, captures the denial reason and reference number, and posts it back into athenaOne so a certified coder can decide the fix. For patients, it calls on aged balances, explains what a colonoscopy statement actually covers, and routes anyone with a billing question or a payment-plan request to the right staff member. Before scheduled procedures it confirms eligibility so a coverage surprise does not show up after prep instructions have already gone out.

Coding accuracy and appeal strategy stay with your people, because that is where training and judgment matter (AAPC). The agent frees those people from hold queues so they spend their day on the claims that actually need a human.

The business case for a procedure practice

Revenue cycle performance comes down to cost to collect and preventable write-offs, and both improve when follow-up is consistent instead of occasional. GI practices have an advantage here: procedure claims are high enough in value that recovering a modest number of aged denials each month pays for the automation with room to spare.

Do the arithmetic for your group. Average denied-claim value multiplied by the number of denials that currently age past the appeal window gives you the annual revenue you are leaving with payers. For a busy endoscopy practice that figure is rarely small, and it exists almost entirely because the follow-up calls never got made.

Administrative by design

The agent handles status calls, balance calls, benefit checks, and routing. The moment a call surfaces something that belongs to a coder or a clinician, it hands off with the details already captured. Money and information move automatically; the decisions that require licensure and training stay with staff.

Key Takeaways

• GI claims fragment into procedure, anesthesia, pathology, and facility charges, each with its own denial path.
• Denials are a structural leak, and most are never appealed simply because no one has time to work them.
• An AI voice agent runs payer status calls, works patient balances, and confirms benefits before procedures, writing back to athenaOne.
• Keep certified coders on coding and appeal decisions; move the repetitive phone-and-portal follow-up to automation.
• Estimate your exposure: average denied-claim value times denials aging past appeal deadlines equals annual dollars left on the table.

GI billing does not break because the rules are too hard for your team. It breaks because the routine follow-up eats the day before anyone gets to the claims worth fighting for. Hand that follow-up to a voice agent and let your billers do the work only they can do.

Sources

• https://www.kff.org/private-insurance/issue-brief/claims-denials-and-appeals-in-aca-marketplace-plans/
• https://www.aapc.com/resources/what-is-medical-billing