Skip to main content

​
Build Overview

This section guides you through how to build, configure, and launch AI voice agents on the platform. It covers the end-to-end process from initial setup to agent deployment, so you can get live quickly and scale with confidence.

​
Why this phase matters

In the Build phase you define how your agents will behave: what they say, how they respond, how they integrate into your campaigns, and how they connect to your data (CRM, leads). A well-designed build leads to higher success rates, better conversational flow, and greater ROI on call campaigns.

​
Key steps in the Build phase

Here are the major steps you’ll follow:
  1. Platform Setup & Branding
    • Configure your account, brand the platform (white-label if applicable), and set up team access and permissions.
    • Define your organizational structure: sub accounts, client accounts, agent groups.
  2. Agent Definition
    • Create an AI voice agent: name it, assign a “personality” (tone, voice, script style).
    • Choose or upload a script/call-flow: define prompts, questions, branches.
    • Upload any custom audio or recordings if you use them for training.
  3. Knowledge & Data Integration
    • Attach a knowledge-base or lead-data source so your agent can refer to real information (e.g., CRM records, client details).
    • Set up CRM integrations and other data-sync tools so the agent can read/write relevant data.
  4. Voice & Language Configuration
    • Select the voice (male/female), accent, and language for the agent.
    • Configure telephony parameters: caller ID, number provisioning, DNC lists, retry logic.
  5. Testing & Preview
    • Use the built-in test or playground mode to simulate calls, check conversation flow, and debug logic.
    • Test edge cases like voicemails, no answer, callback lives, agent hand-over to human.
  6. Campaign Assignment & Launch Readiness
    • Assign the agent to a campaign: upload leads, set time windows, concurrency limits, retries.
    • Verify usage quotas, calling credits, and readiness for outbound/inbound calls.
    • Launch the agent live when ready — monitor initial calls closely for any adjustments.
During the Build phase, favour simplicity in your first agent: choose a short script, one voice, and one direct goal (e.g., “book appointment”). Once you have one working agent, scale to multiple agents, languages and use-cases.