Back to Documentation

Topic Boundaries

Define the operational scope and safety guardrails for your AI assistant.

Guardrails & Safety

To ensure your assistant remains professional and focused on your business, Mentoros implements strict topic boundaries. These guardrails prevent the AI from discussing unrelated topics or providing potentially harmful information.

Configuring Boundaries

Allowed Topics

Explicitly define what the assistant should talk about (e.g., product features, shipping info, company history).

Forbidden Topics

List topics the assistant must avoid (e.g., competitors, political opinions, internal system details).

Fallback Responses

When a user strays outside the defined boundaries, the assistant will provide a polite, branded fallback response. You can customize these messages in theMerchant Portal to maintain a consistent brand voice even when declining to answer.

Was this article helpful?