Artificial intelligence is quickly becoming part of everyday healthcare operations. Practices are using AI to summarize documentation, draft communications, and improve administrative efficiency.

But when it comes to medical coding, one recent experience reminded us of an important lesson:

AI can sound completely confident and still be wrong.

Recently, one of our revenue cycle specialists asked an AI assistant a straightforward coding question.

The response came back instantly.

The AI confidently identified CPT® 37225 as the correct code for the documented procedure.

There was just one problem.

CPT® 37225 had been deleted effective December 31, 2025.

Fortunately, the error was identified before the code could be used.

This wasn’t a hypothetical example.

It happened during normal coding research and serves as a reminder that AI should support healthcare professionals, not replace them.

Figure 1. Microsoft Copilot incorrectly identified CPT® 37225 as a current and valid code in 2026. The code had been deleted effective December 31, 2025.

Educational Note: AI continues to evolve and can be an excellent productivity tool. However, coding recommendations should always be verified using current coding resources, payer guidance, and qualified coding professionals before claims are submitted.

Why This Matters

Healthcare coding changes continuously.

Each year, CPT® codes are revised, deleted, and added. Medicare Administrative Contractors (MACs), commercial insurance carriers, and other regulatory organizations also update billing policies, documentation requirements, and reimbursement guidelines throughout the year.

While AI systems are becoming increasingly capable, they may rely on outdated training data, incomplete references, or older coding information. Even when an answer is presented with confidence, that doesn’t guarantee it reflects the most current coding guidance.

For physician practices, relying solely on AI for coding decisions can introduce unnecessary financial and compliance risk.

The Financial Impact of Coding Errors

Using an outdated or deleted CPT code can lead to:

• Claim rejections
• Payment delays
• Increased denial rates
• Additional staff time spent correcting claims
• Delayed cash flow
• Increased compliance risk

For busy physician practices, even one incorrect code can create unnecessary administrative work, delay reimbursement, and increase operational costs.

Where AI Adds Real Value

This experience isn’t an argument against artificial intelligence.

In fact, AI is already improving many areas of healthcare operations.

It can help teams:

• Summarize documentation
• Organize information
• Draft emails and communications
• Analyze trends
• Improve administrative efficiency
• Assist with research

These capabilities can save valuable time.

The key is understanding where AI ends and professional judgment begins.

Why Human Expertise Still Matters

Medical coding isn’t simply matching a diagnosis with a procedure code.

Accurate coding requires understanding:

• Current CPT® code revisions
• Medicare billing requirements
• Commercial payer policies
• National Correct Coding Initiative (NCCI) edits
• Modifier usage
• Medical necessity requirements
• Documentation standards

These decisions require current knowledge, experience, and clinical context that AI cannot consistently verify on its own.

The Future Isn’t AI or Humans. It’s Both.

Artificial intelligence will continue transforming healthcare. We believe that’s a positive development.

The most successful physician practices won’t replace experienced professionals with AI. They’ll combine the efficiency of AI with knowledgeable coding and revenue cycle experts who verify accuracy, understand payer requirements, and recognize when something doesn’t look right.

Technology can make your team faster. Human expertise helps keep your revenue protected.

Key Takeaway

Artificial intelligence is an excellent tool.

It is not a substitute for current coding expertise.

As this real-world example demonstrates, AI can confidently provide outdated coding guidance that could lead to claim denials, delayed reimbursement, or unnecessary compliance risk if not verified.

The best revenue cycle combines technology with experienced professionals who know how to identify problems before they impact your bottom line.

Frequently Asked Questions

Can AI accurately assign CPT codes?

AI can assist with coding research and administrative tasks, but coding recommendations should always be verified using current coding resources and qualified coding professionals.

Why should physician practices verify AI-generated coding recommendations?

Medical coding changes every year. AI may reference outdated CPT® codes, payer policies, or billing guidance if its information is not current.

What happens if a deleted CPT code is submitted?

Submitting a deleted CPT code may result in claim rejection, delayed reimbursement, additional administrative work, and potential compliance concerns.

Is AI replacing medical coders?

No. AI is becoming a valuable productivity tool, but experienced coding professionals remain essential for interpreting documentation, applying payer rules, ensuring compliance, and protecting revenue.

About Peregrine Healthcare

Founded in 2001, Peregrine Healthcare partners with physician practices nationwide to strengthen revenue cycle performance through medical billing, provider credentialing, payer contracting, front-office support, compliance consulting, and operational expertise.

We believe technology should enhance healthcare operations, but protecting your revenue still requires experienced professionals who understand today’s complex reimbursement environment.

Sources

• American Medical Association (AMA). CPT® Professional 2026.
• Centers for Medicare & Medicaid Services (CMS). Medicare Claims Processing Manual.
• CMS National Correct Coding Initiative (NCCI) Policy Manual.