HCE 128: Private Behavioral Intelligence

​

HCE 128 for Fraud

HCE 128 turns prior, privacy-safe customer activity into a compact behavioral fingerprint, helping surface transactions and sequences that are unusual for that customer. A separate client-local Context Graph adds approved relationship context. HCE supports fraud decisions.

​

HCE 128 for Private AI Infrastructure Security

An emerging, customer-controlled application for AI model-release and cloud-native environments. It correlates identity, GitHub, CI/CD, cloud, Kubernetes, and runtime signals to surface behavioral escalation. Core scoring requires no public LLM and is designed to operate without customer-data egress. Initial design-partner evaluations are private, shadow-only, and non-enforcing.