On March 26, 2026, OpenEvidence launched Coding Intelligence — and in doing so, quietly redefined what a clinical evidence platform is for.
OpenEvidence was built as an evidence search engine. You asked a clinical question, it searched peer-reviewed literature, and it returned a cited, grounded answer. It was the tool clinicians used to check an unfamiliar treatment, verify an emerging guideline, or settle a clinical argument with evidence rather than opinion.
Coding Intelligence is something fundamentally different. It delivers automatic ICD-10 diagnosis suggestions derived from the clinical documentation, E/M level recommendations with the full MDM (medical decision-making) rationale already written into the note, CPT code suggestions for procedures and services performed during the visit, and automatic CPT code sequencing to maximise reimbursement under Medicare's Multiple Procedure Payment Reduction rules.
This is not an evidence feature. It is a revenue feature. And the fact that OpenEvidence — the company that positioned itself as the purest play in medical evidence search — is now building billing tools tells you something important about where the clinical AI market is heading.
What Coding Intelligence Actually Does
The product operates within OpenEvidence Visits, the company's clinical documentation and communications suite. After a consultation is documented, Coding Intelligence analyses the note and generates coding recommendations.
ICD-10 suggestions are surfaced automatically, reflecting the actual complexity of what was documented — not a simplified code selected from habit or memory. The suggestions are designed to capture the full clinical picture, including secondary diagnoses and comorbidities that many clinicians under-report.
E/M level recommendations come with the MDM rationale already written. For every hour of patient care, physicians spend nearly two additional hours on documentation, and MDM is one of the most time-consuming components. Coding Intelligence generates the MDM breakdown from the clinical note — whether billing by complexity or by time, the supporting documentation is already done.
CPT suggestions are surfaced based on what was actually performed during the visit, including uncommon procedure codes that are frequently missed for complex cases. OpenEvidence highlights expected RVU values alongside each suggestion so codes can be sequenced correctly before the claim is submitted.
As one early user put it: the algorithm "captures the complexity of the work already being done without forcing the physician to upcode."
Why Now?
Three forces are converging.
First, the AMA's 2026 CPT code set explicitly added AI-related codes across multiple specialties — cardiology, wound assessment, diagnostic analytics. AI services are becoming explicitly billable, which creates both an opportunity and a complexity problem for clinicians.
Second, the documentation-to-revenue pipeline is where the money is. Ambient scribing alone saves time but does not directly increase revenue. Coding — accurate, complete, properly sequenced coding — does. The vendors that connect documentation to reimbursement have a fundamentally stronger business model than those that only connect documentation to the chart.
Third, OpenEvidence's competitors are already there. Abridge publicly foregrounds billable AI-generated notes. Ambience Healthcare has chart-aware coding pilots. Dragon Copilot has announced proactive ICD-10 specificity suggestions. OpenEvidence was the last major clinical AI platform without a revenue-side product. Now it has one.
What This Means for UK Clinicians
Coding Intelligence is a US product, built for US billing systems (ICD-10-CM, CPT, E/M, RVU). It has no direct UK equivalent.
But the strategic signal matters for UK practice too. The direction of travel is clear: clinical AI platforms are expanding from documentation into every adjacent workflow — coding, billing, communications, and follow-up. UK clinicians should expect the same pattern to reach NHS primary care, adapted for QOF, SNOMED CT, and the GP contract rather than CPT and E/M.
For UK clinicians today, the knowledge layer remains the most immediately valuable component of any clinical AI stack. iatroX provides the guideline-grounded clinical reference that supports every clinical decision — regardless of which documentation or coding tools are layered on top. As the full-visit AI stack evolves, the knowledge layer is the constant.
Conclusion
OpenEvidence Coding Intelligence is a market-defining launch. The company that built its reputation on evidence search has moved into billing, revenue optimisation, and coding automation — joining Abridge, Ambience, and Dragon Copilot in the documentation-to-revenue race.
For clinicians, the implication is clear: the tools that document your care are now also the tools that bill for it. The ones that get the coding right will capture revenue that was previously left on the table. And the knowledge layer underneath — the guideline-grounded reference that ensures the clinical decisions are sound before they are coded — becomes more important, not less, as AI takes on more of the billing workflow.