Scoped access
Agents work within defined permissions for the records and tools they need. Read-only analysis is separate from actions that change records or contact suppliers.
Security & controlsHuman authority.
Human authority.
Visible accountability.
Scoped access, explicit approval and an attributable evidence trail govern how agents work with enterprise data.
Operating controls
Clear boundaries around every action.
Explicit approval
Supplier-facing actions and commercial commitments follow the agreed approval workflow. Your designated reviewers see the evidence and proposed action before it proceeds.
Attributable records
Sources, recommendations, approvals and actions remain connected, so your team can trace a decision and identify who authorized it.
Data handling
How we protect your data
AgentrixAI processes your data for the agreed work, with access, retention and handling requirements defined for your deployment.
Encryption
Your data is protected with AES-256 encryption at rest and TLS 1.3 in transit.
Tenant separation
Customer environments are isolated to keep your data separate, with access boundaries defined for your deployment.
AI processing
Your data is not used to train shared AI models. Model providers, processing locations and data-use terms are covered in your agreement.
Retention and deletion
Retention, export and deletion requirements are agreed for your source data, outputs and audit records as part of the deployment scope.
Benchmark participation
A benchmark network is being explored, not offered as a current dataset. Any future participation would be optional and require a separate agreement before data is contributed.
Security documentation
Our team can discuss current security documentation and independent assessment status as part of your vendor review.
Security review
Security support for your team
We work with your security and procurement teams to address access controls, subprocessors, processing locations, incident handling and retention requirements for your deployment.