By Kacie Geretz, Director of Growth Enablement
June 23, 2026
Define and apply a clear coding philosophy that distinguishes meaningful error from minor variance so accuracy reflects actual performance, not scrutiny levels
Ensure your audits evaluate against the same documentation, utilize the same coding philosophies, and apply the correct guidelines to ensure consistency -- regardless of whether work is done by a human or an autonomous coding engine
Continuously monitor autonomous coding performance and use emerging variance patterns as a signal to refine internal guidelines and better configure the engine
The original version of this blog appears on Libman Education’s website here and was published on June 11, 2026.
Most organizations assume their audit process will naturally translate to autonomous coding. In reality, this is where things can become more nuanced.
Autonomous coding does not just change how work is done. It changes how performance is evaluated.
When human coders are reviewed, there is an understood margin for error. When an engine is doing the coding, the expectation often shifts toward a much higher standard. That shift is not necessarily a bad thing. In many ways, it should be expected as part of the value of the engine. The goal is better, more consistent, and more scalable performance. But that higher standard needs to be applied thoughtfully and with an understanding of how an autonomous medical coding engine actually functions.
There is a difference between errors that materially impact reimbursement or compliance and variances that are more subjective, for example, whether a code should have been unspecified versus other specified, with no real impact on the outcome. If those distinctions, or rather what I call coding philosophies, are not clearly defined, it becomes very easy for accuracy measurements to reflect scrutiny levels rather than actual performance.
This is where alignment and defining your coding philosophy become critical. Coding philosophy means how your organization defines what constitutes a meaningful error versus a minor variance and applies that standard consistently, regardless of whether the work was done by a human or an engine.
If a pattern of variances starts to emerge, that is not just an audit finding; it is a signal. It may point to an opportunity to refine internal coding guidelines or to better configure the autonomous coding engine to reflect your organization’s coding philosophy. I’ve seen teams spend weeks chasing secondary diagnosis capture when the real issue was the underlying MEAT or TAMPER criteria not being properly configured within the engine.
The practical way to support consistency regardless of who is performing the coding is straightforward. Blind your audits. Auditors should not know whether they are reviewing human-coded or engine-coded charts. Accuracy should always be evaluated against the same documentation, coding philosophies, and guidelines. This keeps your accuracy and audit data comparable over time and ensures you are measuring performance, not perception.
It is also important to recognize that autonomous coding performance is not static. These models evolve over time. With ongoing updates, retraining, and exposure to broader datasets, performance can improve, but it can also shift in ways that are not always favorable.
Depending on the technology and development approach, in some cases, model updates influenced by data from other organizations can introduce changes that affect your environment. This is where ongoing monitoring becomes essential. Accuracy is not something you validate once at go-live. It needs to be continuously tracked on a consistent basis so you can quickly identify changes, understand their impact, and adjust as needed.
At the end of the day, the expectation should be clear. Autonomous coding should deliver better, more consistent results over time. But getting there requires upfront alignment, ongoing oversight, and a willingness to continuously refine both your audit approach and the technology itself.
Check out this Peer Insights Playbook for insight into how Intermountain Health and Ohio State University Physicians approached coding accuracy and auditing after going live with Nym's autonomous coding engine.
Kacie Geretz, RHIA, CPMA, CPC, CCA is the Director of Growth Enablement at Nym, where she aligns Nym’s product roadmap with the evolving needs of health system partners and serves as the externally-facing expert on Nym’s autonomous medical coding engine. A graduate of The Ohio State University’s Health Information Management program, Kacie brings deep expertise across the revenue cycle—having led revenue integrity programs, built managed care contracting and credentialing infrastructure, and driven denials and A/R process improvement initiatives. She is passionate about advancing healthcare automation and regularly shares insights on coding innovation and RCM transformation.
.png)