Security
Your relationships are your advantage. We treat them that way.
Human approval for AI actions
CRM notes, scheduled posts and outbound messages always pause for a person. Approved actions run exactly once.
Policy per action, not per prompt
What the agent may do is fixed for each action. Instructions hidden in imported content are treated as data and flagged.
Encrypted credentials
Integration tokens are encrypted at rest and stored separately from their metadata. Tokens are redacted from every log.
Verified webhooks
Incoming webhooks are signature-checked, time-boxed and protected against replay.
Secure sign-in to integrations
OAuth connections use state and PKCE, with verifiers kept server-side.
Roles and seats
Owner, admin, operator, member and viewer roles control who can approve, connect and configure.
Audit log
Approvals, permission changes, integration changes, repairs and replays are recorded with who and when.
Organization-scoped data model
Every record belongs to an organization, and the graph is partitioned by organization.
Have a security questionnaire?
Book a call and we'll walk your team through how GROW AI handles data, access and AI actions.