The Most Overlooked Feature with AI Scribes: Coding Awareness

As a practicing physician and former chief innovation officer, I experienced firsthand how the growing number of administrative tasks are leading to clinician burnout and becoming a major barrier to direct patient care. 

The thoughtful application of artificial intelligence (AI) in healthcare has the potential to completely transform how care is delivered in the United States. Among its most transformative applications is the AI medical scribe, which saw major adoption in 2024. 

By listening in real-time to patient-clinician interactions and automatically generating detailed clinical documentation, AI scribes are now saving clinicians hours of paperwork, reducing burnout, and allowing more time for direct patient care. 

However, as health systems rush to deploy these technologies, one critical feature is often overlooked — coding awareness..

What is coding awareness?

As physicians, we are trained to focus on delivering the best clinical care for our patients.

We are not necessarily trained on how our documentation drives downstream processes. Evaluation and management coding, DRG assurance, and HCC scoring are not part of our educational rubric. These processes, however, are essential to the financial integrity of an institution, the quality and risk adjustment of populations, and the driving of life-saving research, disease registries, and publicly reported outcomes.

AI platforms must be trained to support the nuances of critical downstream processes.  

First defined by the American Academy of Professional Coders (AAPC), coding awareness is the ability of an AI medical scribe to navigate the labyrinth of complex medical billing and coding systems and ensure that the documentation it generates is compliant with coding and billing rules. This involves more than just a superficial familiarity with billing codes. A truly coding-aware AI must:

1. Comprehend the nuances of thousands of codes – From ICD-10 and CPT codes to the ever-expanding list of modifiers, medical billing is a detailed and dynamic field. An effective AI scribe must understand these intricacies to structure documentation appropriately.
2. Incorporate distinct logic for specific codes – Every medical code has its own logic and rules. For example, coding for a particular diagnosis might require documentation of a symptom with specific language, treatment plan, or follow-up recommendation. AI scribes must be able to apply these rules accurately.
3. Use precise, code-compliant language – The phrasing in clinical documentation must reflect the specific language required by billing codes. Ambiguity or imprecision can result in rejected claims or compliance issues. In certain billing arrangements, the output of clinical documentation must reflect the specific documentation required by certain diagnoses codes. HCC capture requires MEAT-compliant documentation.

Without coding awareness, AI scribes can — and do — produce documentation that is incomplete, non-compliant, or inconsistent with the selected billing codes. This oversight is more than a technical hiccup; it’s a systemic risk.

The risks of “coding naive” AI scribes

I’ve met with health system leaders across the country who have deployed “coding naive” medical scribes. At the beginning, these organizations experienced a major increase in clinician satisfaction and significant decreases in documentation time. However, as their deployments continued, they saw massive increases in Clinical Documentation Integrity (CDI) queries when documentation that was generated did not support selected codes. This results in major downstream issues for the organization. 

AI scribes that lack coding awareness introduce significant challenges for healthcare providers and administrators. When documentation doesn’t align with billing codes, it gets flagged by CDI and revenue cycle management teams. This leads to:

1. Increased administrative burden – Instead of alleviating the administrative load, a poorly designed AI scribe can exacerbate it. Clinicians and administrative staff are left to correct errors or fill in gaps manually, negating the intended efficiency gains.
2. Compliance risks – Incorrect or incomplete documentation can lead to audits and claim denials. For health systems already struggling with tight margins, these risks are unacceptable.
3. Inaccurate billing and problem list adjudication – Documentation that does not accurately and thoroughly capture the encounter results in underbilling, whether it be reduction in E/M codes, or missing add-on CPT codes that were in fact substantiated by the care provided.
4. Delays in patient care – Documentation discrepancies may require clinicians to revisit notes or clarify details after the fact, delaying billing and potentially delaying follow-up care for patients.

The importance of coding awareness

Coding-aware AI scribes provide solutions to these challenges by embedding CDI knowledge directly into their systems. As AI tools are always accountable to humans, this technology allows both clinicians and professional coders to operate at their top of license. This proactive approach saves time, reduces administrative burdens, and supports better financial outcomes for healthcare organizations.

The path forward

To fully leverage the potential of AI medical scribes, healthcare organizations must prioritize coding awareness in their adoption strategies. This includes evaluating AI tools based on their ability to:

1. Support accurate coding decisions at the point of care
2. Proactively prompt if a selected code is not supported by documentation
3. Provide comprehensive compliance with payer requirements
4. Demonstrate proven impact on CDI metrics

AI scribes that lack coding awareness risk undoing years of CDI progress, increasing claim denials, and creating new administrative challenges. In contrast, coding-aware AI scribes can become indispensable tools for improving workflow efficiency, enhancing reimbursement accuracy, and ultimately delivering better patient care.

By insisting on coding-aware AI solutions, health systems can ensure that these technologies fulfill their transformative promise, reducing clinician burnout and administrative burden, improving patient care and communication, and accurately capturing what occurred during an encounter. The journey to smarter healthcare must prioritize compliance, precision, and process integrity as foundational elements of AI deployment. 

Furthermore, any deployment of AI must always be accountable to the clinicians and healthcare workers that it was built to support, elevating and enabling their work. Clinical documentation serves a multitude of healthcare professionals, and so should our AI platforms.

Photo: FG Trade, Getty Images

This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.