AKASA Debuts Autonomous AI for Inpatient Coding and Documentation

The mid-cycle represents the most resource-intensive phase of hospital administration, where clinical records are translated into billing codes. Historically, this process has relied on manual labor, with industry data indicating error rates as high as 20%. By fine-tuning generative AI models to the specific documentation habits and patient demographics of individual health systems, AKASA claims its platform can match or exceed human performance in key metrics, including MS-DRG assignment and principal diagnosis accuracy.

For hospitals, the shift from human-led review to automation offers a significant compression of timelines. While a typical inpatient encounter requires 30 to 60 minutes of manual coding, AKASA’s system completes the task in less than 90 seconds post-discharge. This acceleration is intended to reduce accounts receivable days and mitigate the impact of staffing constraints. The platform is currently being integrated into operations at major institutions, including the Cleveland Clinic, which is exploring the tool to improve precision in its complex patient workflows.

AKASA’s expansion comes as the company continues to scale its footprint, with its customer base now accounting for approximately 10% of all U.S. inpatient discharges and over $180 billion in aggregate net patient revenue. Following the rollout of inpatient coding and clinical documentation integrity (CDI) modules, the company plans to extend these autonomous capabilities to outpatient facility encounters in the near future.