Skip to main content

What you’ll see

How to build, configure, and deploy AI voice agents for both inbound and outbound calls — covering agent setup, voice selection, call-behaviour settings, and how to review performance once the calls start rolling in.

Your agent library

  • Browse every agent in a card or table layout
  • Search by name or filter by type, status, or tags
  • See performance at a glance — total, successful, and failed calls, plus last active time

Building an AI agent

  • Pick an LLM model and dial in its creativity with a temperature setting
  • Choose from 12+ voices across Cartesia, ElevenLabs, and Deepgram
  • Set a speech-to-text provider for live transcription
  • Write your own system prompt — or auto-generate one — to define the agent’s personality and goals

Fine-tuning call behaviour

  • Call duration — cap each call (defaults to 5 minutes)
  • Voicemail detection — skip calls that aren’t answered live
  • Silence threshold — control the natural pacing of conversation
  • Background ambience — add subtle room sound for a more natural feel
  • Recording & transcription — capture and transcribe calls live

Welcome message options

  • Silent start — wait for the contact to speak first
  • Dynamic — the AI opens based on the contact’s context
  • Defined — use a fixed, scripted opening line

Routing & outbound

  • Assign a dedicated inbound and outbound agent to any phone number
  • Launch outbound calls manually from a contact record or via workflow automation
  • Pass live contact data to the agent mid-call
  • Check credits and call cost before every call

Call logs & analytics

  • Full history: date, duration, cost, and call type
  • Outcomes marked success or failure, with the disconnection reason
  • Sentiment analysis on every call
  • Improvement suggestions ranked High, Medium, or Low priority