AI Medical Coding Companies: The 2026 List
Who automates medical coding in 2026, how autonomous each platform is, what they report on accuracy, and which fit your care setting. Updated October 2026.
The Short Answer
The leading AI medical coding companies in 2026 are Fathom, CodaMetrix, Nym, and AKASA for autonomous hospital and physician coding, with Epic now offering native autonomous coding for radiology and emergency medicine, Solventum and Optum for computer-assisted coding, and Lime Health AI for AI-assisted ICD-10 coding in home health and hospice. The dividing line is autonomy: fully autonomous engines suit high-volume routine claims, while payment-linked settings like PDGM favor AI drafting that your credentialed coders review.
The 2026 AI Coding Market at a Glance
Reported positioning as of October 2026. Confirm current claims and pricing with each vendor.
| Company | Built for | Model | Reported accuracy / traction |
|---|---|---|---|
| Lime Health AI | Home health & hospice (PDGM) | AI suggests codes and flags errors for your coders | 95% documentation accuracy, 30,000+ charts/yr |
| Fathom | Health systems, emergency medicine | Autonomous | Reports 90%+ of charts coded end-to-end |
| CodaMetrix | Health systems (multi-specialty) | Autonomous (CMX platform) | Mass General Brigham spinout, No. 1 in 2026 Best in KLAS for autonomous coding |
| Nym | Health systems (ED, radiology) | Autonomous (Clinical Language Understanding) | Epic Toolbox designation for fully autonomous coding |
| Solventum 360 Encompass | Hospitals (HIM departments) | Computer-assisted coding | Category incumbent, formerly 3M HIS. Solventum announced plans in August 2026 to separate the business |
| Optum | Hospitals & payers | Computer-assisted coding, enterprise RCM | Largest RCM vendor footprint |
| AGS Health | Health systems | Hybrid AI + offshore coding services | Large services base adopting AI tooling |
| RapidClaims | Practices & RCM teams | AI coding + denials automation | Newer entrant, RCM-focused |
| Ambience AutoCDI | Health systems using Ambience scribe | Coding inside an ambient documentation suite | Bundled with ambient notes, no human sign-off |
| SimiTree SARA | Home health (outsourced review) | AI-assisted OASIS and coding review services | AI layer on an outsourcing model |
| Epic native coding | Epic health systems (radiology, emergency medicine) | Autonomous, built into the EHR | Announced at Epic UGM, August 2026. Low-confidence charts route to review |
| AKASA | Health systems (inpatient) | Autonomous inpatient coding and CDI | Launched October 2026. Cleveland Clinic and Nebraska Methodist named as customers |
| Arintra | Health systems (multi-specialty) | Autonomous coding plus CDI and denial appeals | $25M Series B (August 2026), $51M raised in total |
How to Read This Market in 2026
AI coding has split by autonomy level. Autonomous coders (Fathom, Nym, CodaMetrix, AKASA, and since August 2026 Epic itself) target high-volume hospital specialties where most charts follow patterns: emergency, radiology, pathology. Routine charts get coded and submitted with no human touch, and coders handle exceptions. Computer-assisted coding (Solventum, Optum) remains the hospital HIM standard: the software suggests, humans confirm. Verified AI coding keeps a credentialed human on every chart while AI does the drafting, the model built for settings where a coding error changes the payment grouping.
Home health sits firmly in that last category. Under PDGM, the primary diagnosis and comorbidity codes place each 30-day period into one of 432 payment groups, coding depends on OASIS-E2 assessment context rather than a single encounter note, and CMS targets the category with ADRs and audits. That is why Lime's ICD-10 coding draws AI suggestions from the full visit recording and flags missing or potentially incorrect codes for your own coders, and why the autonomous hospital engines don't compete in this setting.
What Changed Since August 2026
Four developments reshaped the hospital side of this market in two months:
- Epic entered autonomous coding. At its August 2026 user group meeting, Epic said its native autonomous coding is available for radiology and emergency medicine, with surgery and pathology planned. High-confidence charts are coded automatically and low-confidence charts are marked for review. Fathom, Nym, and CodaMetrix all sell into Epic health systems, so Epic customers now compare third-party engines against a built-in option.
- Solventum plans to separate its Health Information Systems business. Solventum announced on August 5, 2026 that it will pursue a separation of the business behind 360 Encompass, which it says generates $1.4 billion in annual sales and is used by more than 75% of U.S. hospitals.
- Arintra raised a $25M Series B in August 2026, led by Define Ventures, for autonomous coding plus documentation and denial-appeal automation.
- AKASA launched autonomous inpatient coding and CDI on October 2, 2026, naming Cleveland Clinic and Nebraska Methodist Health System as customers.
None of these products target home health or hospice. PDGM coding still depends on OASIS-E2 context that hospital engines do not read.
Sources: Healthcare IT Today (Epic UGM 2026), Solventum, Arintra, AKASA, CodaMetrix.
Software, Services, or Both
Agencies evaluating coding help are usually choosing between outsourced coding services (McBee, Corridor, SimiTree, HPS) and AI platforms. The service vendors are adding AI layers (SimiTree's SARA), and AI platforms like Lime help in-house coders work faster. The practical questions: who is accountable for each chart, what does it cost per chart, and does the workflow live inside your EMR. We compared the main options in our home health coding and QA services comparison, and the build-vs-outsource math in Lime vs outsourced coding.
FAQ
AI Medical Coding, Answered
What coding managers and agency leaders ask when evaluating this market.
An AI medical coding company builds software that reads clinical documentation and assigns billing codes (ICD-10, CPT, HCC) automatically. The market splits into autonomous coders that submit codes with no human touch on routine charts (Fathom, Nym, CodaMetrix), computer-assisted coding that suggests codes for human coders to confirm (Solventum, Optum), and setting-specific AI coding where AI suggests codes and flags errors for the agency's own coders (Lime Health AI in post-acute care).
Vendor figures are self-reported, so third-party signals help: CodaMetrix ranked No. 1 in the first Best in KLAS segment for autonomous coding (2026), Fathom reports automating over 90% of charts end-to-end, and Nym earned Epic's Toolbox designation for fully autonomous coding. AKASA, which launched autonomous inpatient coding in October 2026, cites blinded third-party testing against human coders. Accuracy varies by specialty and chart complexity. In payment-linked settings like PDGM home health, Lime Health AI suggests ICD-10 codes from clinician notes and flags missing or potentially incorrect codes, so your own coders can review faster.
Autonomous coding vendors typically price per chart coded or per encounter, negotiated on volume. Computer-assisted coding platforms price per seat or as enterprise licenses. Coding services price per chart reviewed. Lime Health AI uses per-chart volume pricing for its AI coding and OASIS review software, with pilots available and price matching.
The autonomous coders (Fathom, Nym, CodaMetrix, AKASA, and Epic's own native coding) are built for hospital and physician-practice claims, not post-acute care. Home health coding is different: ICD-10 selection drives the PDGM payment grouping, depends on OASIS-E2 assessment context, and gets audited by CMS accordingly. Lime Health AI is purpose-built for this: AI suggests codes from the full visit context and flags missing or potentially incorrect codes, so your coders can correct them before they reach the EMR. SimiTree's SARA also applies AI to OASIS and coding review as part of its outsourcing services.
In high-volume, pattern-heavy specialties like radiology and emergency medicine, autonomous coding already handles most routine charts, and coders shift to exceptions and audits. In context-heavy, payment-linked settings like home health, the durable model is AI plus human judgment: AI does the drafting and error checks at scale, and the agency's credentialed coders apply clinical judgment. Lime builds software for that model, to make your coders faster, not to replace them.
Home health or hospice agency? See AI coding suggestions built for PDGM.
Request a Demo