Security & Governance.

Vyasa implements role-based access control (RBAC) across the platform. Approval matrices route exceptions by amount, category, and risk tier, so the right person sees the right decision at the right time.

• Role-based access control for operators, approvers, and administrators
• Approval matrix configurable by amount thresholds and workflow type
• Agent actions scoped to policy-permitted operations only

Every agent decision is logged with a full reasoning trace: the policies applied, the precedents consulted, and the evidence attached. Decision lineage is stored in the context graph, queryable for compliance and post-hoc review, not reconstructed from email.

• Immutable audit log for all agent and human actions
• Decision graph with timestamped lineage
• Export-ready audit reports for compliance teams

Vyasa is designed with human-in-the-loop as a core principle. Agents recommend and execute routine decisions; high-stakes actions always route to human approvers with full context, simulation, and precedents. Vyasa does not train on customer data.

For full AI disclosure and usage policies, see our AI Disclosure page.

Data in transit is encrypted using TLS 1.2+. Data at rest is encrypted using AES-256. Vyasa operates with a zero data storage model, acting across your existing systems without copying business records to vendor cloud infrastructure.

Vyasa is actively working toward SOC 2 Type II certification. We maintain honest status on compliance milestones and will update this page as certifications are achieved.