AKASA Expands Autonomous AI Platform Into Inpatient Medical Coding and CDI

AKASA has expanded its autonomous AI platform into inpatient medical coding and Clinical Documentation Integrity (CDI) — the most complex and highest-stakes layer of hospital billing — and is claiming human-level accuracy on the cases where a single classification error can swing reimbursement by thousands of dollars.

For most health systems, inpatient coding is a staffed specialty function, not an administrative task. Certified coders and CDI nurses review clinical documentation after every discharge, assign Diagnosis-Related Group (DRG) codes, and flag documentation gaps before claims are submitted to Medicare, Medicaid, and commercial payers. Getting it right determines how much the hospital gets paid. Getting it wrong means underpayment, clawback audits, or compliance exposure. AKASA is now saying a machine can handle all of it, autonomously, at scale.

What launched

According to HISTalk, the expanded platform automates the full mid-cycle revenue workflow. From the moment a patient is discharged, the system reviews clinical documentation, assigns inpatient codes, checks CDI gaps, and queues corrections — without routing cases to a specialist coder or CDI nurse for initial review.

AKASA's previous platform focused on prebill review: catching errors after a human had already coded the case. The new expansion moves the automation earlier. The system does the coding first, not the coder.

Key claims from the launch:

• Market: $12 billion revenue cycle management industry
• Accuracy claim: Human-level performance across complex inpatient DRG categories
• Scope: Autonomous coverage from discharge to clean claim submission

Why hospitals are listening

Hospital operating margins have been under sustained pressure for three years. Labor costs are up. Reimbursement rates haven't kept pace. And the revenue cycle workforce — coders, CDI nurses, billing specialists — is facing the same recruitment and retention crisis as clinical staff, but with less visibility.

Certified Professional Coders (CPCs) take three to six months to train, earn competitive salaries, and are increasingly difficult to retain as demand from remote-work opportunities grows. Coding backlogs translate directly to delayed cash collection. For a health system processing 4,000 inpatient discharges a month, a two-week backlog is a material cash flow problem, not an operational inconvenience.

AKASA's proposition is that autonomous AI eliminates the backlog problem entirely — processing discharges in near-real-time regardless of case volume or staff availability.

Why this is harder than claims automation

Prior authorization and claims status checking — the two categories where revenue cycle AI already has significant adoption — are largely rule-based tasks. Inpatient coding is not. Complex DRG assignment requires interpreting free-text clinical notes, understanding complication and comorbidity (CC/MCC) hierarchies, sequencing principal diagnoses correctly under Official Guidelines for Coding and Reporting, and applying Medicare Severity adjustments accurately. It's a clinical reasoning task with billing consequences.

AKASA's human-level accuracy claim will get intense scrutiny from health system buyers and their compliance teams. Health systems are directly liable for coding accuracy — Medicare's Recovery Audit Contractors (RACs) conduct post-payment audits and pursue clawbacks on improperly coded claims. "The AI did it" is not a recognized defense in a RAC audit.

What to watch

Independent audits of AKASA's accuracy across diverse patient populations, payer mixes, and facility types will determine whether the clinical-accuracy claims hold beyond the company's own case studies. Health systems evaluating the platform will also need to resolve workforce transition questions upfront: what happens to existing coding staff, and how does coding liability flow when an AI system is the original coder of record.

The $12 billion revenue cycle management market is one of the few segments of healthcare AI where the ROI math runs directly to the CFO's income statement. Expect competitive responses from established RCM vendors including Optum, nThrive, and Ciox Health.