Security
Security is a feature, not a PDF.
How we're different
- Execution over tracking: we do close work, not just manage tasks.
- Inbox and docs automation with long-running follow-ups.
- Audit-first with immutable action logs and evidence links.
Trust rule: AI drafts, humans decide for sensitive outcomes.
Controls
- Encryption in transit and at rest
- RBAC + org roles
- Immutable audit logs
- Least-privilege connector scopes
- Data retention controls
- Safe mode on uncertainty
AI reliability
- Structured outputs (JSON schema)
- Deterministic validation for numbers and dates
- Approval gates for sensitive actions
Not legal advice, not a substitute for your compliance obligations.
Audit Pack Score (example)
92/100
- Tasks complete: 95%
- Evidence attached: 90%
- Approvals recorded: 100%
- Open exceptions: 2
Illustrative. In the MVP console, the score is computed from workflow state.
Trust center
Security, privacy, reliability, and model usage boundaries in one place.