Skip to main content
Corti code prediction models convert unstructured clinical text into structured medical codes. On their own, they return predictions — codes the model is confident should be billed, candidates worth human review, evidences pointing back to the source text, and alternatives the model considered. Building a production coding workflow means combining those predictions with your own context: payer contracts, billing requirements, EHR data, and reviewer tooling. Corti agents can further enrich the pipeline by querying external knowledge or running multi-step reasoning over the note before or after coding. These guides walk through the core implementation patterns, from foundational encounter coding to CDI review and revenue cycle automation.

What’s covered

Encounter Diagnosis Coding

Assign accurate diagnosis and procedure codes to clinical encounters — the foundation workflow.

Clinical Documentation Integrity (CDI)

Accelerate concurrent and retrospective CDI review by surfacing code suggestions and documentation gaps.

Revenue Cycle

Surface missed codes and revenue leakage at scale with HCC capture sweeps and retrospective audits.

Where to start

New to the Medical Coding API? Start with Encounter Diagnosis Coding — it covers the foundational request patterns, review workflows, and success metrics that the other guides build on.

Common coding system combinations

  • US inpatient admission: icd10cm-inpatient + icd10pcs
  • US outpatient / ED / office visit: icd10cm-outpatient + cpt
  • UK inpatient admission: icd10uk-inpatient + opcs4
  • UK outpatient: icd10uk-outpatient
  • German inpatient admission: icd10gm-inpatient + ops
  • German outpatient: icd10gm-outpatient
  • French inpatient admission (PMSI): cim10fr-inpatient + ccam
  • French outpatient / ambulatory: cim10fr-outpatient + ccam
  • International inpatient: icd10int-inpatient (or a national extension such as icd10gm-inpatient)

Please contact us if you need help choosing the right workflow or have questions about implementation.