Skip to content

Analytics workflows

Analytics workflows let you analyze and transform encrypted data without ever decrypting it. They provide a secure environment for data exploration, feature engineering, and model training using Python or PySpark—while ensuring that all computations remain confidential and policy-governed.

Unlike agentic workflows, which stay live and respond to requests programmatically, analytics workflows run as discrete jobs. Each job executes once within a trusted enclave, producing results that can be shared, audited, and verified without exposing the underlying data.

Analytics workflows are ideal for privacy-sensitive analysis, regulatory reporting, and machine learning over protected datasets—where maintaining confidentiality is as important as generating insight.

In this section