PII Redaction (Prompt-Time)

Prompt-time PII redaction is the removal or substitution of personally identifiable information from an AI prompt before it leaves the user's network, so the model and its provider never receive the original values.

Most enterprise AI deployments rely on user training, contractual data-handling clauses, or post-hoc log scanning to manage PII exposure. None of these prevent the values from reaching the model provider.

Prompt-time redaction is a hard technical control. It runs in the request path before egress and either blocks, warns, or substitutes — the values that triggered the rule never appear in the outbound HTTP request.

For HIPAA, GLBA, GDPR, and CCPA workloads, prompt-time redaction is the difference between a defensible compliance posture and an audit finding. It is also the only control that protects against accidental paste of regulated data into a chat box.