ICD-10 coding with AI Clinical Notes

How AirNote supports ICD-10 coding

Guides

For many US therapists, ICD-10 coding is an essential part of clinical documentation. The right code supports accurate records, communication with other providers and, where applicable, insurance claims and reimbursement.

It can also add another administrative step to an already busy day.

What are ICD-10 codes?

ICD-10-CM codes are standardised codes used in the United States to classify diagnoses and health conditions. They provide a shared clinical language: for example, distinguishing between different anxiety, depressive, trauma-related and adjustment disorders.

The system helps therapists, healthcare providers, insurers and record systems refer to a condition consistently. The CDC describes ICD-10-CM as the standardised system used to code diseases and medical conditions in the United States.  

Why accurate coding matters

An ICD-10 code may influence whether a claim is accepted, how medical necessity is represented and how a client’s diagnosis appears across their clinical record.

Selecting a code therefore requires care. A code should reflect the clinician’s assessment and documented diagnosis—not simply a symptom mentioned during the session.

At the same time, finding and checking the most appropriate code can interrupt the documentation workflow. Similar diagnoses may have closely related codes, while contextual factors can sometimes be represented with additional Z-codes.

Why AI is a good fit for coding support

ICD-10 recommendation is a useful role for AI because it involves reviewing structured clinical material, recognising relevant terminology and comparing that information with an established classification system.

AI can quickly:

  • Surface a likely code for consideration.

  • Check that the code and description correspond.

  • Explain which documented information informed the suggestion.

  • Reduce the need to search manually through code lists.

  • Apply the same careful process across every note.

The appropriate role of AI, however, is to support coding—not determine a diagnosis.

Clinical meaning is often nuanced. Symptoms alone may not establish a psychiatric disorder, and the therapist may hold information that is not present in a particular session transcript. The final decision must therefore remain with the clinician.

How AirNote supports ICD-10 coding

Therapists can enable ICD-10 Codes in AirNote’s settings. Once enabled, AirNote can include an ICD-10-CM draft recommendation when it generates a Clinical Note.

AirNote uses a curated set of codes relevant to therapy and mental health practice. It reviews the documented session material, checks whether there is an adequate clinical basis for a suggestion and validates the proposed code before displaying it.

Each recommendation is clearly presented as a recommendation for clinician review, together with its supporting rationale. AirNote does not treat symptoms alone as sufficient evidence for a psychiatric diagnosis, and it will not force a recommendation when the available information is too limited or uncertain.

In those circumstances, the safer result is simply: no recommended code suggested.

Less searching, with the therapist still in control

AirNote’s ICD-10 functionality is designed to remove a repetitive part of documentation while preserving clinical accountability.

It does not diagnose the client or submit a billing-ready code automatically. Instead, it gives the therapist a carefully checked starting point within the Clinical Note, making it easier to confirm, revise or reject the suggestion.

The principle is straightforward:

AirNote recommends. The therapist decides.

Used in this way, AI can make ICD-10 coding faster and more consistent—without replacing the professional judgement on which accurate diagnosis and documentation depend.