AgentLink Feature: Delegated AI Governance

The demo shows how developers can expose real SaaS tools to AI in minutes, enforce global guardrails, and now empower their customers with granular tenant-level policy control. Policies can include approvals, spending limits, time and region restrictions, and other dynamic conditions that help organizations safely adopt AI in their workflows.

Key topics:

• Enforcing access control with global guardrails that always apply
• Allowing customers to set their own tenant level constraints for AI usage
• Approvals and step-up authentication for sensitive actions
• Blocking or limiting specific tool calls based on conditions like spend thresholds
• Supporting multiple policy sources that work together at runtime
• Real time policy evaluation that considers user, context, and intent