Building and deploying a voice agent on Vozzo involves a straightforward, linear workflow. This guide breaks down the architecture and lifecycle so you know exactly how the API components fit together.
The Lifecycle
1
Setup & Initialization
You begin by calling /create-agent with your desired language, model (Orbit or Quantum), and voice ID. This returns a persistent agent_uuid.
2
Configuration & Knowledge
Using /update-agent, you inject the agent's personality via agent_prompt. You can then upload PDFs or scrape URLs to build its knowledge base, giving it factual context.
3
Deployment
Attach the agent to a communication channel. Use the Web SDK for sites, or /agent/deploy-phone/ to link a PSTN phone number.
4
Monitoring & Insights
Monitor live interactions using the /agent/agent-conversation/ API to fetch transcripts and analyze call sentiment.
Relationship to Campaigns
Unlike Vozzo Campaigns (which are designed for high-volume, asynchronous outbound dialing), AI Agents are persistent entities usually designed for inbound traffic or web-based support.
However, you can use an AI Agent inside an outbound Campaign by passing its agent_uuid when creating a Campaign. This overrides standard IVR logic with the agent's dynamic LLM.
Agent Dashboard →
Track usage and aggregate metrics
Campaigns →
Learn about high-volume outbound calling