The American Academy of Professional Coders (AAPC) recently released a detailed whitepaper explaining how AI tools used for clinical documentation and coding can either help the revenue cycle or accidentally harm it.
Since many hospitals and clinics are adopting AI tools to reduce documentation time, AAPC decided to outline what actually makes an AI tool useful, and what warning signs providers should watch out for.
The whitepaper explains that while AI tools can save time, some of them produce documentation that misses important details needed for correct coding, medical necessity, and claim approval. When documentation is incomplete, the result can be denials, underpayments, or delays. So the message is simple: AI can improve coding accuracy, but only if the tool is designed the right way.
This blog walks through AAPC’s main points in a simple and easy-to-understand way, and explains how these guidelines relate to real billing work.
Why Accurate Documentation Still Matters
Even with AI helping doctors, documentation still needs to clearly show three basic things:
- What the patient came in for
- What care was given
- Why that care was needed
Insurance companies depend on these details to decide whether to approve or pay a claim. Since rules differ from one payer to another, and change often, it’s hard for busy clinicians to remember everything that must be included.
Because of this, health systems spend years building training programs, improving templates, and creating support teams to help keep documentation accurate. In the same way, professional medical billing companies have developed strong internal processes where trained experts review documentation, coding, and billing to make sure everything is done correctly.
But if an AI tool creates notes that look neat on the surface yet leave out important details, all that effort can fall apart quickly. Clean documentation only works when nothing essential is missing.
Quick stats to help paint the picture:
Due to fear of denials, providers often undercode and lose an average of 5–10%. This usually happens because some part of the process wasn’t handled correctly, and providers worry that using the proper code might cause the insurance company to reject the claim.
Another study shows that, on average, providers spend $43.84 to fight a single denied claim, and at the industry level, this adds up to $19.7 billion in one year.
This is a huge amount. This shows why getting the basics right in documentation, and later in billing and scrubbing, is so important, because one small mistake can lead to a denial from the insurance provider.
Where Basic AI Tools Struggle
AAPC highlights in their AI Awareness whitepaper that many AI tools on the market only create a simple summary of the visit. These summaries may sound clean, but they miss important information needed for coding and billing.
The whitepaper gives examples, which can be explained simply like this:
- In emergency visits: AI left out key information about test results and medical decisions, causing the visit to be coded lower than it should have been.
- For chronic conditions: The AI failed to pick up on the details that show just how severe the patient’s illness actually was, which leads to incorrect risk coding.
- In hospital cases: The tool left out critical treatment information, a mistake that causes huge discrepancies in the final reimbursement.
- For prior authorization: AI failed to record pain levels, disability scores, and treatment goals, all details insurance companies need.
- The pattern is the same across all cases: The AI didn’t understand which details mattered for billing, so the documentation looked fine but wasn’t complete.
Why AI Tools Need Coding Awareness
AAPC explains that an effective AI tool must understand what payers expect. It can’t just summarize the conversation, it must recognize the details that support the codes and medical necessity.
When AI is designed this way, it helps produce more accurate notes, reduces errors, and cuts down on avoidable denials. It also takes a huge weight off the billing team, saving them from having to circle back and fix errors after the fact.
Real-Time Support Helps Reduce Errors
AAPC also highlights the importance of real-time reminders. Since the AI listens to the full visit, it can remind clinicians if something important is missing or unclear. This is helpful because the information is still fresh.
These reminders can help capture:
- Missing details
- More accurate codes
- Additional services provided
- Information that is easily forgotten during a busy day
This reduces the stress on both clinicians and billing teams and keeps documentation complete from the start.
AAPC’s 5-Step Checklist for Choosing the Right AI Tool
AAPC breaks down what an effective AI tool should be able to do, and each step has a clear purpose. Here’s what they recommend looking for:
- The AI should support the clinician’s coding decisions: The tool should be able to capture the details behind the provider’s coding choices. If a certain code is selected, the AI must include the information that proves why it was chosen. This helps prevent claims being questioned later.
- It should warn the clinician if something is missing or incorrect: If a detail doesn’t line up with the code, or something important isn’t documented, the AI should alert the provider right away. This early warning helps avoid simple mistakes that can lead to denials.
- It should document more than basic notes: A basic summary isn’t enough for billing. It is critical that the AI captures the nuances payers require, such as diagnostic interpretations, treatment justification, and future care plans. Without that context, even appropriate care can appear insufficient.
- Certified auditors should vet sample notes before approval: Before the organization fully commits, trained auditing experts need to review a batch of notes to confirm accuracy. This acts as a crucial safeguard to prevent the tool from introducing errors into your documentation.
- The AI tool should be monitored during rollout to track its impact on claims: Even if the AI works well in tests, real workflows can be different. During the rollout phase, teams should watch how the tool affects claims, coding accuracy, and denials. If anything unusual happens, it can be fixed early.
Our Approach to AI Medical Billing and Coding
At Talisman Solutions, we align our process directly with the principles outlined by the AAPC. The AI tools we use are designed to capture the specific details that actually impact coding, documentation, and claim approvals. However, we don’t believe in putting your revenue on autopilot.
We never rely on technology alone. Our billing experts are involved at every step to verify that the work is accurate, compliant, and fully supported by the notes. We let the Ai the volume, while our team handles the quality. That human verification is the difference between a claim looking correct and actually being correct, which saves you from unnecessary delays down the road.
Conclusion
AAPC’s whitepaper shows that the future of clinical documentation won’t come from picking any AI tool, it will come from choosing the right one and using it wisely. When healthcare organizations select tools that understand real coding needs and pair them with trained experts, the entire billing workflow becomes stronger, not harder.
