Security and Data Handling
Designed for practical SMB trust.
Arx-Labs uses customer isolation, permissioned tool access, human review, and clear support-access controls so small teams can use AI teams without losing control of client data.
Customer data
•
Customer data is not used to train AI models.
•
Data is used to provide the customer's AI teams and retain necessary working context during the contract.
•
Data is encrypted in transit and at rest.
•
Customers can request export and deletion according to the data retention policy.
Access controls
•
Agents receive only the permissions required for their workflow.
•
Connector access can be revoked by the customer at any time.
•
Support and admin access is disabled by default.
•
Support access, when needed, is customer-granted, time-limited, and logged.
Human review
•
Arx-Labs AI teams produce drafts, summaries, checklists, and reports.
•
Client-facing messages require human approval unless the customer explicitly configures otherwise in later product stages.
•
Sensitive-domain playbooks require human review by design.
Sensitive domains
Arx-Labs does not provide legal, tax, accounting, payroll, insurance, healthcare, or hiring advice. Playbooks in these areas are for drafting, organization, and informational support only. Qualified professionals remain responsible for final decisions and communication.
•
No professional advice. Outputs require human review before use.
•
Drafting and coordination only. The assistant prepares materials; your team controls final recommendations.
•
Human approval required for all client-facing communication in sensitive domains.
•
Clear boundaries documented per workflow during setup.
Retention
•
During contract: data retained as needed to provide workflow continuity.
•
After termination: export/return window, then deletion by default unless archival or extended retention is selected.
Questions about security or data handling?
We are happy to walk through our controls with your IT lead, compliance contact, or legal counsel.