Skip to main content
AI systems are powerful but not infallible. Even though Stigg’s AI agents are designed for high accuracy, agent misinterpretation is still possible. Stigg understands that some organizations may be concerned about fully delegating the provisioning and invoicing processes to AI agents. To address this, Stigg offers built-in guardrails that help manage risk and ensure control:
  • Support for “Human-in-the-loop” Flows: Human supervision can be achieved either through the Stigg UI — where you review and approve each AI-generated response — or by implementing approval flows programmatically via the APIs or the Workflow Builder.
  • Adjustability: You can edit AI-generated responses as needed, such as adjusting prices, plan names, entitlements, or billing terms before final approval.
  • Traceability: All AI-generated subscriptions and invoices include metadata flags (e.g., createdByStiggAI: true) to ensure full auditability and traceability.
We strongly recommend starting with a semi-automated flow, where a human reviews and approves each AI response. Once you’re confident in the system’s accuracy, you can move toward fully automated workflows.