Home/Handbook/What data trains our models
CLINICAL AI

What data trains our models

We train on a mix of licensed clinical text, synthetic data generated under clinical supervision, and customer data — only when the customer has explicitly opted in and only after deidentification.

We do not train on data from one customer to improve a model serving another, unless both parties have opted in. This is contractual, not aspirational.

Every training data source is logged, with provenance, in our model registry.

Curee AI

Ready to Deploy Your Digital Workforce?

See how Curee Workstation can automate your most time-consuming clinical, operational, and revenue cycle workflows. Live in days.