Trust center
Risk management
Short-form vendor risk answers (internal / illustrative).
How we're different
- Audit exports include integrity verification.
- Supervisor gates for sensitive actions.
- Evidence objects with hashing and provenance.
Trust rule: AI drafts, humans decide for sensitive outcomes.
Short answers
Do you train models on our data?
No, not by default. Model usage is bounded to drafting and summaries; evidence must be deterministic and cited.
Where is our data stored?
Deployment-dependent (local/VPC/on-prem roadmap). Evidence exports support customer-controlled retention strategies.
How do approvals work?
Policy gates require explicit supervisor approvals, captured with rationale and exported with audit proofs.
Do you support audit integrity?
Yes. Audit events are append-only and hash-chained; exports can be verified (API endpoint).