Is AI Coding Correctly?

Corliss
Apr 6
2 min read

Updated: Apr 7

AI in Healthcare RCM is Failing Despite Heavy Investment

The core problem is that AI tools are being built on flawed foundations — trained only on historical claims data without clinical context — and can't keep up with constantly changing payer rules Becker Hospital Review.

We focus on Three Key Failure Points here:

Denials are getting worse, not better. Denial rates jumped from 30% (2022) to 41% (2025), even as AI adoption increased. AI essentially learns and repeats past mistakes. So, you be the judge! Is AI Coding Correctly?
AI Can't Keep Pace with Payers. Payer rules change daily/weekly; most AI models update quarterly. The result: outdated insights, misplaced trust, and missed errors causing revenue leakage.
Pilots Don't Scale. AI is deployed as disconnected point solutions rather than integrated systems. Instead of reducing labor, it's creating more manual work — staff spends more time correcting AI than doing the original task.

Bottom Line:

Is AI Coding Correctly? Not according to 95% of enterprise AI pilots that are falling short. However, the technology isn't failing because AI can't work in RCM — it's failing because it's being implemented inappropriately: wrong training data, wrong architecture, and wrong scope.

So, let’s work on changing this trajectory and focus on getting it right! #RevenueCycle; #HealthcareAI; #HealthTech; #AgenticAI; #P3Quality

About the Author

Corliss Collins, BSHIM, RHIT, CRCR, CCA, CAIMC, CAIP, CSM, CBCS, CPDC, serves as the Principal and Managing AI Advisor of P3 Quality™, a Healthcare Tech company specializing in AI for Revenue Cycle research, development, and issue-resolution management. She also serves as a subject-matter expert and a member of the Volunteer Education Committee at the American Institute of Healthcare Compliance (AIHC). She is a Member of the Professional Women's Network Board (PWN).

Disclosures/Disclaimers:

Is AI Coding Correctly? This analysis draws on research, trends, and innovations in AI for Revenue Cycle Management (RCM). Some of the blog content is generated by AI. Reasonable efforts have been made to ensure the validity of all materials and the consequences of their use. If any copyrighted material has not been appropriately acknowledged, use the contact page to notify us so we can make the necessary updates. LinkedIn Profile: www.linkedin.com/in/ccollinsrhitcca

P3 Quality™