AI Code of Conduct

This AI Code of Conduct establishes standards for deploying AI-powered agents through convobix to ensure ethical, fair, and transparent use of artificial intelligence in customer interactions.

AI Transparency

All AI agents deployed through convobix must clearly identify themselves as AI systems when asked. You must not configure your agent to claim to be human. Your widget configuration should include appropriate disclosure that customers are interacting with an AI assistant.

Accuracy Standards

You are responsible for the accuracy of your knowledge base content. Outdated or incorrect information in your knowledge base may cause the AI to provide inaccurate responses. Review and update your knowledge base regularly. Enable source citations where possible so customers can verify information.

Escalation Requirements

Your AI agent must have a clear escalation path to a human agent for: requests involving significant financial decisions; medical, legal, or safety matters; emotionally distressed customers; and situations the AI is not equipped to handle. The human handoff feature must be functional and monitored.

Bias & Fairness

Ensure your knowledge base and agent instructions do not introduce discriminatory content. Regularly review conversation logs for evidence of biased responses. If you identify systematic bias in AI responses, contact support@convobix.io so we can investigate at the model level.