Data Flows
| Owner | Classification | Version | Effective | Next review | Status |
|---|---|---|---|---|---|
| Sekhar Prakash | Internal | 0.1 | 2026-08-27 | 2026-11-27 | Draft |
Purpose. Documents how customer and personal data moves through the platform across four stages — ingress, inter-service transfer, the model boundary, and egress. Supports data-protection impact assessments and verifies that each flow crossing the trust boundary is controlled and logged.
Scope and method
Four egress points carry sensitive data (per the tokenization design):
- The model prompt.
- The intelligence-log reasoning field.
- Application logs.
- Conversation-checkpoint state.
Each flow in the table below is described by four fields: source, transform (tokenise or encrypt), destination, and retention.
Flows
| Flow | Source | Transform | Destination | State |
|---|---|---|---|---|
| Ingress | Uploads, email, connectors (SAP, Xero, Zoho Books, SharePoint, Salesforce, NetSuite, QuickBooks — tenant-credentialed) | Classification at boundary | Data stores | To be defined |
| Inter-service | Service to service | Tenant scoping, TLS | Data stores | To be defined |
| Model boundary | Tool output / prompt | Tokenise, then de-tokenise on return | External LLM | To be defined |
| Egress | Logs, tracing, exports | Redaction | External systems | To be defined |
Flow diagrams to be embedded during completion.
Actions and owners
To be derived from the tokenization and compliance-readiness findings.
Related
- → AI Governance — Tokenization — the model-boundary control
- → Records of Processing — the data within these flows
- → Architecture — Infrastructure — how the transport is built
Revision history
| Version | Date | Author | Change |
|---|---|---|---|
| 0.1 | 2026-08-27 | Sekhar Prakash | Initial structure. |
Finaisse Internal — Confidential. Not for external distribution.