Corti Launches Symphony AI to Transform Medical Coding
Key Points
- Corti introduced Symphony for Medical Coding, an AI system that treats coding as a reasoning task.
- Built on a peer‑reviewed framework and 1.8 million patient encounters, the system uses four sequential agents mirroring human coders.
- Corti claims up to 25% higher clinical accuracy than models from OpenAI, Anthropic, Amazon, Oracle and Microsoft.
- In a Danish study, Symphony identified three times more suicide attempts than official coding records.
- The system supports both U.S. (ICD‑10‑CM, ICD‑10‑PCS, CPT) and European coding environments without retraining.
- Outputs are auditable, linking each code to supporting clinical evidence and flagging ambiguities for review.
- Symphony is available via API and through the Corti Console, with enterprise and sovereign‑cloud options.
- Corti positions accurate coding as essential for health‑system monitoring, resource allocation and intervention design.
Corti, the Copenhagen‑based health AI company, introduced Symphony for Medical Coding, an agentic system that treats coding as a reasoning task rather than simple labeling. Built on a peer‑reviewed framework and a study of 1.8 million patient encounters, Symphony claims up to 25% higher clinical accuracy than models from OpenAI, Anthropic, Amazon, Oracle and Microsoft. The system uses four sequential agents to extract evidence, navigate the ICD index, validate candidates and reconcile final codes, delivering auditable outputs linked to supporting clinical evidence. Available through an API and integrated with the Corti Console, Symphony operates across U.S. and European coding environments and aims to reduce errors that affect billing, reporting and public health data.
Background and Need
Medical coding converts clinical notes, diagnoses and procedures into standardized alphanumeric codes for billing, reporting and public‑health analysis. In the United States alone, the ICD‑10‑CM system contains 70,000 diagnosis codes, and errors are common, costly and often invisible. Traditional AI approaches treat coding as a classification problem, predicting the most likely code based on historic data, which can struggle as coding guidelines evolve.
Corti’s Approach
Corti developed Symphony for Medical Coding using a peer‑reviewed framework called “Code Like Humans,” presented at EMNLP 2025. The research, based on 1.8 million patient encounters, frames coding as a reasoning task. The system deploys four agents in sequence:
- Evidence extractor – isolates conditions in a clinical note.
- Index navigator – searches the ICD alphabetical index for candidate codes.
- Tabular validator – checks candidates against coding guidelines.
- Code reconciler – sequences and validates the final output.
Each step mirrors the decision process of expert human coders, providing auditable links between assigned codes and the supporting clinical evidence.
Performance Claims
Corti asserts that Symphony outperforms AI models from OpenAI, Anthropic, Amazon, Oracle and Microsoft by up to 25% on clinical accuracy benchmarks. In a peer‑reviewed study of Danish patient data, the system identified three times as many suicide attempts as had been officially coded, uncovering cases present in notes and medication records but missed by time‑pressed coders.
Availability and Integration
Symphony is now available via API and through the Corti Console. It supports both the U.S. coding environment (ICD‑10‑CM for diagnoses, ICD‑10‑PCS and CPT for procedures) and European coding environments without local retraining. The system produces auditable outputs, flags ambiguities for human review, and integrates with the Corti Agentic Framework, supporting A2A and MCP standards. Enterprise and sovereign‑cloud deployments are also offered.
Implications for Healthcare
By improving coding accuracy, Symphony aims to enhance the data layer that health systems rely on for monitoring trends, allocating resources and designing interventions. Accurate coding is presented not as a back‑office cost center but as a critical foundation for health‑system decision‑making.