Trust

Security Overview

Last updated: 2026-06-21

This overview describes the security expectations for a production AI business operating system that stores company, customer, financial, file, billing, and AI-assistant records.

These public terms are implementation-ready product copy for launch preparation. Final legal approval, company details, jurisdiction choices, and support contacts must be confirmed by the owner before public paid launch.

Tenant isolation

Company data must remain isolated by workspace and membership. Routes, exports, files, reports, AI tools, billing records, and administrative actions should require authenticated company access before returning data.

Protected operations

Customer-facing messages, payment capture, estimate approval, data deletion, legal/admin actions, provider setup, and phone-system changes require explicit owner-approved workflows or separate operator controls.

Credential safety

Provider credentials, webhook secrets, API keys, raw payment details, and secret environment values must not be printed, shown in public pages, stored in ordinary business records, or returned through app routes.

Files and exports

Uploaded files and generated exports should remain company-scoped, authenticated, private by default, and protected against unsafe file types, path traversal, cross-company access, and cache exposure.

Monitoring and backup

Production operation requires health checks, audit logs, backup creation, backup verification, restore proof, provider readiness proof, and incident escalation before public launch.