JotAudit checks the note before the payer does

The Centers for Medicare and Medicaid Services (CMS), the American Psychiatric Association (APA), and the major commercial payers each publish the documentation rules their reviewers use. JotAudit runs the note against those rules the same day, not after a reviewer opens the file years later.

Every payer writes down what its reviewers look for. A higher-level medical visit needs a stated reason for the extra time and complexity. A therapy session billed by time needs the minutes written down. A note needs a signature, a diagnosis that matches the code, and a plan that matches the diagnosis. Each payer keeps its own list, the lists are long, and payers update them without much notice.

Most practices never check a note against those lists before the claim goes out. The note gets written, the claim goes out, and the payer's own review happens later, sometimes years later. By then the clinician has moved on and the session is a memory. What comes back is a denial, a repayment demand, or an audit.

How an extrapolation audit works

Some payer audits use extrapolation. A payer can extrapolate: it finds an error rate in a small sample of notes, then applies that rate across every claim billed under the same code and demands the difference back. A few flawed notes in a sample of thirty can turn into a repayment demand covering thousands of claims the reviewer never looked at.

One JotPsych customer, the owner of an eleven-clinician practice, went through exactly that. A commercial payer ran an extrapolation audit against her practice. Working alone, before she found JotPsych, she appealed, and the payer reduced the demanded amount by three quarters. She says she is now "shaking in my boots" about payers running that same kind of audit with AI, at a speed and scale she could not match by hand.

What JotAudit checks

JotAudit runs inside every note. It scores the draft against more than 150 rules built from public payer and regulatory documentation: CMS coverage determinations, APA practice guidelines, and the published requirements of Cigna, Aetna, and Optum. The check runs three times, when the note is first drafted, right before the clinician signs, and again if the note changes afterward. It flags what is missing: a stated rationale for a higher-level visit billed under CPT code 99214, a documented session time for a 90837 therapy visit, or a diagnosis that does not support the code billed. The clinician decides what to do with each flag.

JotPsych built the rule library at the same scale as its note-writing engine: the system has scribed over three million encounters, and the rule library comes from machine-processing thousands of pages of regulator and payer documentation.

Checked from the outside

One practice checked JotAudit's work before buying it. Its owner hired a certified medical auditor, accredited by the American Academy of Professional Coders (AAPC), to review the system as part of the purchase decision. The review held up.

A payer audit is not hypothetical. CMS and the major commercial payers already run them, and more of that review is moving to AI. JotAudit checks the note before the claim goes out, when the clinician can still fix what it finds.

Sources: JotPsych, "JotPsych Debuts JotAudit, Turning Behavioral-Health Documentation Into a Standardized, AI-Driven Audit System," January 21, 2026; JotAudit product page, jotpsych.com/audit.