Revenue Cycle Insider

Understand how existing codes are being fused with new technology.

As artificial intelligence (AI) continues to transform healthcare, gastroenterology practices are increasingly using advanced tools to enhance detection, streamline workflows, and improve patient outcomes. From real-time polyp detection during colonoscopies to automated coding suggestions obtained from clinical documentation, AI is transforming both the clinical and administrative sides of GI care. For medical coders, this evolution brings new opportunities — as well as new challenges.

Jessica Miller-Dobbs, CPC, CPC-P, CGIC, during her AAPC HEALTHCON 2026 presentation “The Bowel of Truth: Tackling Coding, Denials, and Mapping the End-to-End Revenue Cycle for GI,” set out to give a clear understanding as to how these emerging technologies integrate with existing coding frameworks, documentation requirements, and payer expectations.

Continue reading to understand how to better prepare yourself for AI merging with your existing GI practice.

Visualize AI in Your Practice

AI is playing a huge role in gastroenterology practices in the way of enhancing computer-aided detection by visually outlining lesions during live procedures and helping the endoscopist focus on potential cancer polyps or inflammatory bowel disease (IBD). “AI is now used to detect abnormalities in thousands of images with high accuracy and assisting in efficient review of massive data sets,” said Miller-Dobbs.

She then went on to describe the importance of many types of AI-related technologies being used in GI practices across the United States:

• Optical character recognition (OCR) takes scanned-in colonoscopies, pathologies, and imaging reports and turns them into actionable electronic health record (EHR) data. Doing so creates a higher accuracy in data extraction than ever before at 95 percent. This will help with clinical trial recruitment, quality reporting, and by reducing manual and administrative workload. Miller-Dobbs said, “AI is reducing the physician burnout we’ve been seeing for the last 15 or 20 years.”

According to Miller-Dobbs, this technology is currently being used to scan through hundreds of thousands of medical records and endoscopy videos to identify certain patients matching inclusion criteria for clinical trials, with Crohn’s disease and ulcerative colitis being the focus at the moment. OCR is being used to expedite the process of prior authorizations, thereby reducing the workload for all involved and decreasing claim processing time.

• Natural language processing (NLP) reads and understands human language and turns notes, reports, and EHR text into structured, usable data. “AI is reading physician notes and automatically suggesting your ICD-10-CM codes, and can even flag for missing documentation,” said Miller-Dobbs.
• GI Genius™ (Medtronic) is the first U.S. Food and Drug Administration (FDA) approved technology to provide real-time computer-aided detection (CADe) for polyps during a colonoscopy procedure. The technology will overlay a green box on the video monitor whenever a potential polyp is found.
• TissueCypher is an AI-driven precision medicine test that analyzes biopsied tissue to predict the risk of progression from Barrett’s esophagus to esophageal cancer and provides a personalized five-year risk assessment (low, intermediate, high) to guide treatment decisions. “This technology uses spatial context of cells and tissue systems to be preserved while investigating morphology and multiple cellular measures. It’s looking for the biomarkers and spatial relationships within various cell populations and tissues,” she explained. Keep in mind that TissueCypher is not FDA approved at this time.

Understand How Existing Codes Work With AI

According to Miller-Dobbs, as of the publication of this article, the codes listed below are still used for normal and AI-assisted procedures, but you will need to pay attention to the practitioner’s notes and be sure it is clearly stated to the payer if AI was used during a procedure:

Miller-Dobbs also advised documenting the clinical decision-making for the payer. Here’s one example she gave to the audience: “Last week, AI highlighted a 4 mm lesion in the sigmoid colon. What this physician did was make the comment that ‘Upon direct visualization, this was confirmed to be an essential polyp and was removed by smear polypectomy.’ This not only helps the AI, but it shows that you have reviewed what the AI flagged, and protects you from an audit.”

Avoid Documentation Red Flags

When the practitioner performs AI-assisted procedures, it’s important to avoid being audited for poor documentation. Here is a helpful list of documentation red flags she provided:

• AI is mentioned, but no physician review is documented
  • Claim may be denied; potential false attestation 
• Vague note: “AI-assisted colonoscopy performed”
  • Insufficient for audit defense; the physician needs to add specifics
• More polyps were removed than were documented
  • Upcoding risk — the count must match pathology 
• AI module isn’t functioning, yet it was billed as AI-assisted
  • Fraudulent claim; billing must match equipment records
• Copy-forward/clone documentation
  • All notes must be specific to each encounter.

Lindsey Bush, BA, MA, CPC, Production Editor, AAPC