AKASA brings autonomous AI to hospital coding and clinical documentation – Unite.AI

AKASA on October 2, 2026 announced the launch of an autonomous AI platform for mid-cycle healthcare revenue, expanding the South San Francisco company from AI-driven prebill review to autonomous hospital medical coding and clinical documentation integrity (CDI).

AKASA, which describes itself as a leader in generative AI for the healthcare revenue cycle, said its customers account for more than $180 billion in aggregate net patient revenue and about 10% of hospital patient discharges in the U.S., about 1 in 10. The company said the volume of hospitalized patients processed through its AI products grew nearly 6-fold in the year prior to launch, crediting those products with the expansion of its healthcare system’s customer base.

AKASA said it refines AI models for individual health systems, taking into account differences in patient populations, clinical criteria, documentation practices and complexity of care, an approach the company says supports comprehensive documentation and accurate coding that better reflects the care provided. CEO and co-founder Malinka Walaliyadde considered the launch an industry milestone. “For years, a self-contained midcycle has been the holy grail in our industry,” he said. “Today AKASA is making it real.”

Because mid-cycle automation was difficult

The intermediate cycle is the phase of the revenue cycle in which a patient’s medical record is translated from documentation to codes that determine reimbursement, quality reporting, risk adjustment and patient record integrity. The ad describes the work as highly complex and resource-intensive, and still predominantly manual.

This complexity, the company said, made reliable automation difficult. The announcement references a 2025 peer-reviewed study by npj Health Systems that reported medical coding error rates of up to 20% and a July 2026 U.S. Government Accountability Office review that identified verifiable accuracy as a central challenge for healthcare organizations adopting AI for medical notes and coding.

According to the company, it typically takes a coder 30 to 60 minutes to code a single hospital encounter, and workforce shortages and capacity constraints can leave health systems waiting several days after a patient’s discharge before a coder even begins using the account. AKASA product materials peg the current discharge-to-encode interval at three to four days, versus its system’s 90 seconds.

How the autonomous coding platform works

According to AKASA’s autonomous hospital coding documentation, the system reads the patient’s complete medical record (discharge summaries, operative notes, progress notes, consultations, labs, imaging and medications) and manages each encounter from start to finish, assigning the entire set of codes independently across all specialties. Each code links to the exact graph language that supports it, with the underlying logic and security, which the company says creates an audit trail and leaves validation taking seconds.

Health systems establish operational boundaries. Customers define thresholds, service line rules, payer mix, and audit sampling, and the system codes only what the organization has approved to code, with ICD-10-CM/PCS integrated. The platform works alongside existing electronic healthcare documentation and billing systems, so the move to self-service does not require their replacement. Coding occurs at discharge, when the chart closes; AKASA says this reduces discharged-not-final-coded (DNFC) volume, reduces accounts receivable days, and speeds up cash collection without adding staff.

AKASA said it tested the system through blinded third-party evaluations in which its artificial intelligence and expert human coders coded the same inpatient encounters, which account for 65% to 85% of inpatient volume in most health systems. Independent reviewers, who were not told whether each set of codes came from AI or a human, found that the AI ​​matched or exceeded expert coders on key measures of accuracy, the company reported: MS-DRG assignment, principal diagnosis, clinical quality capture and accuracy of presentation at admission. AKASA said the AI ​​completes coding in less than 90 seconds after discharge and markets the product as the first standalone solution built specifically for hospital coding.

CDI, distribution and timely response

In addition to coding, AKASA is extending the platform upstream into CDI, describing a unified AI layer between clinical documentation, coding and pre-invoice review intended to maintain complete documentation. The company said that during the mid-cycle autonomous rollout it will work with health systems on customized rollout plans to scale autonomous volume as needed, and that it will soon introduce outpatient encounters to the platform.

Cleveland Clinic has used AKASA’s prebill review products across coding and CDI and plans to explore mid-cycle standalone offerings, according to the announcement. “Our revenue cycle work is particularly time-consuming because we care for many medically complex patients,” said Rohit Chandra, chief digital officer at the Cleveland Clinic. “With autonomous coding, we seek to improve speed and accuracy in these demanding processes with a compliance-focused approach to this work.”

Jeff Francis, chief financial officer and vice president of finance at Nebraska Methodist Health System, who has worked with AKASA for several years, described autonomy as the natural next step for health systems, emphasizing revenue integrity, denials, write-offs and how quickly his organization gets paid. Julie Yoo, general partner at Andreessen Horowitz, said healthcare cannot meet the scale of future demand through human labor alone, and characterized AKASA’s technology as frontier AI built for the specific needs of complex clinical documentation and hospital coding.

AKASA said its custom models are trained on each health system’s clinical documentation, case mix and coding decisions, and that its solutions range from artificial intelligence that supports CDI and coding teams to fully autonomous workflows that the code encounters from start to finish.