Governance
Governance and responsible AI
How AXO governs AI output: every answer cited to a source document or refused, permission-aware retrieval, full lineage, and a human accountable for every decision.
The rule the platform is built around
Every answer is cited to a source document, or refused. A model that cannot point at evidence does not get to speak. This is enforced in the retrieval layer, not left to prompt wording.
Human accountability
- AI drafts; a named person approves
- Approvals are recorded against the individual, not the system
- Any output can be traced to the inputs and the version of the model that produced it
Data protection
- Tenant isolation between customers
- Encryption in transit and at rest
- Role-based access control, applied to retrieval as well as to the interface
- Customer-controlled retention and deletion
- Customer data is never used to train shared models
- In-country data residency available — see deployment options
What this page does not claim
AXO does not advertise certifications, audit reports or service-level figures it cannot currently evidence. If your procurement process requires them, raise it early and we will tell you plainly what exists today and what does not.