Summary
Up to 19 percent of E/M visits are undercoded nationwide, at roughly $37 each.1 HFMA puts total coding-related loss at 1 to 5 percent of annual revenue, most of it undercoding.2
Upcoding invites an audit. Undercoding just bleeds, and nobody files a complaint about being paid too little, so it is the error your practice never checks for.
Coding attached at the point the decision was made, drafted against the MDM criteria.
Book a demoHow to tell
Three signals. Your E/M distribution skews hard toward level 3. Your revenue per encounter sits below specialty peers. And your documentation supports a higher level than what you billed.
The specific test for family practice is 99215 utilisation as a share of established visits. Benchmark is 22 to 28 percent.33Family practice benchmark for 99215 utilisation is 22 to 28 percent of established visits. Below 18 percent indicates undercoding or payer downcoding. Below 18 percent and something is wrong, and a submitted-versus-paid variance report tells you whether it is you or the payer.
Why it happens
Coders code what is documented. If the note does not evidence the complexity, the visit gets coded down, and the coder is right to do it. Missing secondary diagnoses are the same story. A comorbidity that affects management has to be captured to reflect the complexity, and a coder reading an incomplete note cannot invent it.
From 2026 the E/M level turns on Medical Decision Making.44From 2026, Medical Decision Making is the primary basis for E/M level selection. Level selection turns on number and complexity of problems, amount and complexity of data reviewed, and risk of complications. Number and complexity of problems, amount and complexity of data reviewed, risk of complications. A physician who has not retrained their documentation habits around those three is exposed on every encounter.
None of this is a coding problem. It is a documentation problem that shows up in the coding.
What WA\ does
WA\ Clinician drafts against the MDM criteria and attaches suggested coding at the point the decision was made. The secondary diagnosis that affects management is in the note because it was in the consultation, and the risk is stated because the risk was real.
You review and sign. The coding is a suggestion and the responsibility is yours, on our product and on every other product in this category.
Against 41 seconds
Take a physician at 4,000 encounters a year. Nineteen percent undercoded at $37 is roughly $28,000. That is one error type, on Medicare rates, before the payer's downcoding starts.
The best randomised result any ambient scribe has produced is 41 seconds a note, and neither product tested reduced after-hours work at all.55Lukac PJ, Turner W, Vangala S, et al. Ambient AI Scribes in Clinical Practice, A Randomized Trial. NEJM AI 2025;2(12). The category is competing on the seconds and the money is in the code.
What we are not claiming
Suggested coding is not a coder and it is not a CDI programme. We do not audit your claims, we do not know your payer contracts, and a certified coder reviewing your distribution will find things we will not.
Upcoding is the other failure and it is worse. A tool that pushes your levels up is selling you an audit. The correct target is the level the documentation supports, which is sometimes lower than what you billed.
Availability
WA\ Clinician runs a 14-day free trial with no minimum term. Pricing is on one page.
Frequently asked questions
How do I know if my practice is undercoding?
Three signals. Your E/M visits skew heavily toward level 3. Your average revenue per encounter is below specialty peers. And your documentation consistently supports a higher level than what was billed. For family practice the specific test is 99215 utilisation as a share of established visits, where the benchmark is 22 to 28 percent and below 18 percent means something is wrong. A submitted-versus-paid variance report tells you whether it is your coding or payer downcoding. A formal coding audit is the most reliable way to confirm and quantify it.
How much does undercoding cost a practice?
AAPC Audit Services reports up to 19 percent of E/M visits are undercoded nationwide at roughly $37 or more each in lost Medicare reimbursement. For a physician at 4,000 encounters a year that is roughly $28,000 from one error type alone, before payer downcoding is counted. HFMA estimates organisations lose 1 to 5 percent of annual revenue to incorrect or incomplete coding, much of it undercoding where legitimate reimbursement is simply never collected. It goes unaudited because nobody files a complaint about being paid too little.
Can AI fix undercoding?
It can address the cause, which is documentation rather than coding. Coders code what is documented, so if the note does not evidence the complexity the visit gets coded down and the coder is right to do it. From 2026 E/M level selection turns on Medical Decision Making, meaning number and complexity of problems, amount and complexity of data reviewed, and risk of complications. A tool that drafts against those criteria at the point the decision was made closes the gap. A tool that pushes your levels up regardless is selling you an audit, and you review and sign either way.
The documentation is the defence.
Start the trial and look at your own E/M distribution after a fortnight. If it has not moved, you have lost two weeks and nothing else.
About this article. Written and published by WA\, which sells a product that suggests coding, so treat the framing accordingly. AAPC, HFMA and benchmark figures are third-party and cited in the margin. Undercoding rates vary by specialty, payer mix and region. Suggested coding is a suggestion and the clinician is responsible for what is submitted. None of this is billing, legal or compliance advice, and a certified coder should review your own distribution before you change anything.