A source submits a value
The collection flow receives a field in the context of a tenant, data type and registered purpose.
Product / Field-level data protection
Give each sensitive field a token that applications can use. When an AI workload needs the original value, MemberGo checks its purpose, identity and field policy before release.
Protection model
Follow a temperature reading from a production line. The token and value shown here are synthetic.
The collection flow receives a field in the context of a tenant, data type and registered purpose.
The vault encrypts the value. Applications link records with a token. Possessing that token grants no authority to retrieve the field.
An approved quality-audit job may use the field in memory and clears it after use. Logs and usage records exclude the raw value.
Tokenization controls where raw values travel. Envelope encryption protects those values in the vault.
Interactive example
The process parameter and service account stay the same. Only the declared purpose changes.
Workload request
pii_7f3a…
Access granted
access_id: demo_ac9dThe approved field is used in memory for this workload. A successful access is traceable. Its original value stays out of logs.
No fields released
403 data_access_deniedThe gateway stops the dependent job. The response reports the denial. Detailed reasons remain inside VISCOR.
Policy design
A tenant registers custom types such as process parameters or study codes and sets allowed purposes, retention and workload access.
Each workload uses a service account in its tenant. Its grant can include model versions, permitted hours and a daily access limit.
MemberGo checks ownership, state, type and processing basis for every field. If any field fails its checks, the whole request is denied.
Release follows successful access and audit writes. Reviewers can trace an access ID to its purpose, workload and model version.
Scope the first evaluation around one workload, one permitted purpose and one case that should be denied.