Vocanta deploys AI voice agents that speak with your customers. This policy describes how those agents are configured and constrained.
Transparency
Where required by law, agents identify themselves as AI at the start of the conversation.
Grounding
Agents answer from the knowledge base and integrations you configure. Out-of-scope questions escalate to a human handoff or a documented fallback.
Training
Customer content processed by your agents is not used to train foundation models. Model improvements are performed on synthetic and licensed corpora.
Human oversight
Customers can review transcripts, tune prompts, and update workflows at any time.