On July 29, 2026, Ebix Health announced the A.D.A.M. Medical Coding Platform, an automated system that maps health content to ICD-10 and LOINC codes at scale. The announcement is worth a second look from anyone in medical coding or RCM, not because Ebix is a claims-coding vendor — it isn’t — but because of what the platform’s design choices say about where automated coding, as a category of software, is heading across the industry.
What Ebix actually launched
Per Ebix’s press release, A.D.A.M. is described as an automated coding system “supported by expert review” that maps health content — patient education material, CME programming, and related clinical text — to ICD-10 and LOINC. The company says it processes thousands of articles daily and produces coded output structured for electronic health records, patient portals, and digital health platforms. The platform sits alongside Ebix’s existing A.D.A.M. patient education library and Oakstone CME business, and the release frames it as an extension of the company’s Health Content and Wellness portfolio rather than a standalone product line. Ebix CEO Gagan Sethi framed the launch around that existing content business, saying the new capabilities “strengthen a business that already has deep clinical credibility.”
Content coding is not claims coding
It’s important to be precise about what this is. A.D.A.M. codes content — educational articles, media, reference material — not clinical encounters or claims. That’s a fundamentally different problem than assigning CPT and ICD-10-CM codes to a physician’s documentation for reimbursement, where sequencing, medical necessity determinations, payer-specific edits, and audit exposure all matter in ways they simply don’t for a patient-education article about, say, managing type 2 diabetes. Ebix’s own framing supports this reading: the stated target audience is content partners, health systems, and digital health platforms building coded libraries, not RCM departments submitting claims for payment.
Why the distinction still matters for coders
Even so, the underlying technical claim — automated mapping of unstructured text to standardized code sets at high volume, with expert human review functioning as a check rather than the primary mechanism — is the same basic architecture increasingly used in claims-side autonomous coding tools. Vendors serving content libraries and vendors serving revenue cycle departments are converging on a similar pattern: run automation first on the bulk of the volume, then route exceptions and edge cases to human coders for review. Watching how that pattern performs and is marketed in a comparatively lower-stakes domain like content coding is a reasonable, if imperfect, proxy for how the same pitch is likely to show up in claims-side coding tools over the next year.
Code set breadth as a signal
One detail worth noting on its own: A.D.A.M. maps to both ICD-10 and LOINC in the same platform, rather than just one code set. That pairing — diagnosis coding alongside laboratory and clinical observation coding — points toward vendors building for interoperability across EHRs, patient portals, and payer systems, since LOINC is what typically carries lab and observation data between those systems while ICD-10 carries diagnosis data. Coding teams evaluating any automated coding vendor, content-side or claims-side, should ask which code sets are actually supported natively versus bolted on, because that answer usually reveals how broadly the vendor intends the tool to be used.
What this signals for claims-side automated coding
A handful of things are worth tracking as more vendors describe products as “automated coding,” regardless of which side of the content-versus-claims line they actually sit on:
- Scale claims need context. “Thousands of articles daily” is a meaningful throughput number for content coding, but it doesn’t translate directly into claims volume, denial rates, or coding accuracy on complex, multi-diagnosis encounters.
- “Expert review” placement matters. Whether human review happens on every output, a sampled subset, or only on flagged exceptions changes the actual audit-risk profile substantially, even when two vendors both use the phrase “expert-reviewed.”
- Code set breadth is a marketing signal. Supporting ICD-10 and LOINC together, as Ebix does, is a deliberate interoperability bet, and it’s worth asking the same question of any coding automation vendor pitching claims-side work.
- Global applicability claims deserve scrutiny. Ebix notes ICD is used worldwide and LOINC is recognized in 30-plus countries, but also flags that local coding conventions and language requirements still apply — the same caveat should hold for any vendor claiming cross-market reach for claims coding.
What coding and compliance teams should do with this
For working medical coders and compliance leads, the practical takeaway isn’t about Ebix specifically — it’s a reminder to keep asking the same questions of every “automated coding” vendor pitch that lands in an inbox this year. What exactly is being coded: content, encounters, or claims? Where does human review actually sit in the workflow, and on what percentage of output? And what’s the evidence for accuracy on the coding tasks that carry real audit and reimbursement risk, as opposed to throughput metrics on lower-stakes content? Announcements like Ebix’s are useful precisely because they’re explicit about scope, which makes those distinctions easier to see than they are in pitches that blur content coding and claims coding under one general “automated medical coding” label.
That distinction is exactly where Medikode’s automated medical coding platform is built to hold up: automation designed specifically for claims-side coding work — ICD-10-CM, CPT, and HCC coding tied to actual clinical encounters — where accuracy, audit defensibility, and reimbursement are on the line, not just content throughput.