Lesson 3: Platform Overview: Build → Test → Deploy → Monitor

Platform Overview

Retell is organized around four stages:

:hammer: Build

  • Agent creation — define personality, instructions, and behavior via prompts
  • Conversation Flows — visual node-based editor for complex multi-step calls
  • Knowledge Base — upload docs, URLs, or text for the agent to reference
  • Custom Functions — call external APIs during conversations (check availability, look up records, etc.)

:test_tube: Test

  • Playground — web-based environment to have test conversations with your agent
  • Simulation Testing — run automated test scenarios at scale to validate behavior

:rocket: Deploy

  • Phone Numbers — buy numbers directly or bring your own (Twilio, Vonage, custom SIP)
  • Inbound/Outbound — handle incoming calls or make outgoing calls via API
  • Batch Calls — launch campaigns of outbound calls
  • Web/Chat — embed agents in websites as voice or text chat

:magnifying_glass_tilted_left: Monitor

  • Call Logs — full transcripts, recordings, and metadata for every call
  • Post-Call Analysis — extract structured data from calls (sentiment, outcomes, custom fields)
  • Webhooks — get real-time notifications for call events
  • AI QA — automated quality scoring across call cohorts