Beyond the Hype: What AI Can’t Fix in CDI and Coding
NLP systems have been parsing clinical notes since the early 2000s. What’s different today is the accessibility and sophistication of large language models, which can suggest queries, generate draft codes, and even simulate training scenarios. Still, these tools can’t erase long-standing people, processes, and policy challenges.
-------
Artificial intelligence (AI) has become one of the most talked-about tools in healthcare. From clinical note summarization to coding assistance, its potential seems limitless. But those of us who have worked in coding, clinical documentation integrity (CDI), and health information management (HIM) know something important: AI is not new, and it’s not a cure-all.
Natural Language Processing (NLP) systems have been parsing clinical notes since the early 2000s. What’s different today is the accessibility and sophistication of large language models (LLMs), which can suggest queries, generate draft codes, and even simulate training scenarios. Still, these tools can’t erase long-standing people, process, and policy challenges.
This article explores what AI can enhance in CDI and coding — and, more importantly, what it can’t fix.
Where AI Enhances CDI and Coding
Smarter Document Parsing
Modern NLP can extract clinical concepts with greater accuracy than early auto-coding systems, helping highlight missing or implied diagnoses. Still, oversight is essential — machines don’t always grasp clinical nuance.
Faster Query Drafting
AI can create compliant draft queries in seconds, cutting down on repetitive work. But it’s CDI professionals who decide whether to send, edit, or withhold those queries.
On-Demand Learning Tools
LLM-powered simulations allow coders and CDI staff to practice without touching PHI. They can rehearse query drafting, test documentation scenarios, and build confidence in a low-risk environment.
Cross-Referencing Guidelines
AI can pull logic from multiple sources — ICD-10-CM, CPT®, HCC, Coding Clinics — and even explain pathways. But coders’ clinical reasoning and sequencing expertise remain central.
Accelerating Reviews
AI is excellent at flagging missing specificity (laterality, severity, causality). What it can’t do is explain why the omission matters to patient care or reimbursement — that’s where human judgment is indispensable.
What AI Cannot Fix
1. A Poor Documentation Culture
If providers aren’t documenting clearly, AI won’t magically rewrite records. Education, communication, and clinician engagement are human-led initiatives.
2. Inconsistent or Non-Compliant Query Practice
AI might generate a query, but if your team lacks standardization in tone, format, or escalation, compliance risks remain.
3. Gaps in Coding Knowledge
AI outputs are only as safe as the people reviewing them. Without strong knowledge of anatomy, physiology, and coding conventions, coders may misinterpret suggestions.
4. Toxic or Fear-Based Leadership
Teams afraid to ask questions or try new tools will not thrive, no matter how advanced the technology. Culture change is built through leadership, not algorithms.
5. Overreliance on Shortcuts
Efficiency without strategy is dangerous. Overusing AI to “speed things up” without clear policies can lead to denials, audit risk, and compliance issues.
The Evolving Role of CDI and Coding Professionals
As AI takes over repetitive tasks, the value of human expertise grows. Coders and CDI specialists will spend more time on:
- Interpreting ambiguous documentation
- Educating providers
- Leading quality reviews
- Advising on compliance and audit readiness
- Strategically integrating AI outputs into workflow
This shift requires not technical expertise, but adaptive thinking, clinical insight, and digital confidence. These skills are becoming core competencies for tomorrow’s workforce.
Scenario & Prompt Practice
One of the safest ways to experiment with AI in CDI and coding is by using synthetic, non-PHI case scenarios. This allows staff to practice query drafting and documentation review without compliance concerns.
Example Scenario
A 65-year-old woman presents for evaluation of uncontrolled hypertension. Her chart lists “Type 2 diabetes” in the problem list and includes an A1c result of 9.0%. The provider notes “neuropathy symptoms” but does not clearly link them to diabetes. Her medication list includes both an antihypertensive and insulin.
Practice Prompt
“Act as a clinical documentation integrity specialist. Review the following documentation and identify potential areas requiring provider clarification. For each gap, explain why clarification is needed, referencing CDI best practices and coding guidelines. Suggest a compliant, non-leading query that could be sent to the provider.”
How to Use It
- See if the AI recognizes the need to clarify the linkage between diabetes and neuropathy.
- Review whether the AI suggests query language that is neutral and compliant.
- Check if the AI properly separates coexisting conditions from causality assumptions.
By running variations — adding CKD, removing the A1c, or noting medication changes — you can test how AI responses shift and where human expertise must override.
Moving Forward: Human Strengths Still Win
AI in CDI and coding isn’t a revolution — it’s the next step in a 20-year journey. These tools can accelerate documentation review, draft queries, and support education. But they can’t replace clinical reasoning, compliance judgment, or organizational culture.
As healthcare embraces automation, CDI and coding professionals must focus on the skills machines cannot replicate: curiosity, integrity, and the ability to interpret context. AI can be a powerful partner, but only if guided by experienced hands.
The takeaway? Don’t rush to replace — instead, integrate thoughtfully. Build governance, train your staff, and let AI enhance (not dictate) your workflows.
Why Partnering with Experts Still Matters
AI may be advancing quickly, but the foundation of compliant, audit-ready documentation and coding remains human expertise. Technology can highlight gaps and generate drafts, but only trained professionals can ensure accuracy, compliance, and defensibility. That’s why many organizations turn to trusted revenue cycle management partners for support.
At Bristol Healthcare Services, our certified coding and CDI specialists bring deep clinical knowledge, regulatory expertise, and proven best practices to every engagement. We help healthcare organizations:
- Strengthen documentation quality and provider education
- Improve coding accuracy and risk-adjustment compliance
- Streamline query workflows with compliant, standardized practices
- Prepare for audits with confidence and transparency
Whether you’re looking to integrate AI tools responsibly or simply need reliable medical coding services and CDI support, our team provides the human expertise that technology alone can’t deliver. Together, we ensure your documentation tells the full clinical story — accurately, compliantly, and with measurable financial impact.