Deploying a Voice Agent
Step-by-step guide to deploying a new AI voice agent for a client, from template selection to live phone number.
Deploy a Voice Agent for Your Client
Deploying a voice agent is the core action you'll take on the platform. This guide covers the full process from start to finish, including what happens behind the scenes and what to do after your agent goes live.
Before You Begin
Make sure you have the following ready:
- An active client profile — you need to create the client in the system first. See Managing Clients if you haven't added them yet.
- Business details on hand — the client's business name, phone, hours, services, and address. You'll enter these during deployment.
- Sufficient Agent Fuel — deploying an agent doesn't consume credits, but the agent needs credits to handle calls once it's live. Check your balance on the Dashboard.
Note: Each client can have one voice agent. If a client already has an agent deployed, you'll see the option to reconfigure or redeploy instead of a fresh deployment.
Deployment Steps
Navigate to the Client Page
Go to Clients in the sidebar and click on the client you want to deploy an agent for. On their Overview tab, click the "Deploy Agent" button.
Browse the Template Marketplace
The Template Marketplace opens, showing 145 pre-built templates across 19 industry categories. You can:
- Browse by category — Automotive, Dental, Legal, Home Services, Health & Wellness, and more
- Search by keyword — type "plumber" or "law firm" to find relevant templates
- Preview templates — click any template to see its description, included prompts, and sample conversation flow
Choose the template that best matches your client's industry and use case.
Fill In Template Variables
Each template has customizable variables tailored to that industry. For example, a plumbing template might ask for:
| Variable | Example Value |
|---|---|
| Business Name | Johnson & Sons Plumbing |
| Business Phone | (555) 987-6543 |
| Service Area | Greater Denver Metro |
| Hours | Monday–Saturday, 7 AM to 7 PM |
| Emergency Service | Yes, 24/7 emergency line available |
| Services | Drain cleaning, pipe repair, water heater installation, sewer line inspection |
Fill in every variable for the best results. The AI uses these details to answer caller questions naturally and accurately.
Select a Voice
Browse the voice library to choose the voice that will represent your client's business. You can filter voices by:
- Gender — male, female, or neutral
- Accent — American, British, Australian, and more
- Style — warm, professional, energetic, calm
Click the play button on any voice to hear a sample. Pick the voice that feels right for the brand — a law firm might want a calm, authoritative voice, while an auto detailing shop might prefer something upbeat and friendly.
Review Your Configuration
Before deploying, review a summary of everything you've configured: the template, all variable values, and the selected voice. This is your chance to catch any typos or missing information.
Click Deploy
Hit the "Deploy" button to launch the agent. The deployment process takes about 15–30 seconds.
Test Your Agent
Once deployment completes, the agent's new phone number appears on screen. Call it right away to test. Have a full conversation — ask about hours, services, location, and appointment booking. Make sure the agent responds accurately with the information you provided.
Success: Your voice agent is now live! It will answer calls 24/7, handle customer inquiries using the business details you provided, and log every conversation with full transcripts and analytics.
What Happens Behind the Scenes
When you click Deploy, the platform performs several actions automatically:
- Phone number provisioning — a dedicated phone number is provisioned through Twilio and assigned to this client's agent.
- AI agent creation — an ElevenLabs Conversational AI agent is created and configured with your selected voice, template prompts, and business information.
- Knowledge injection — all template variables, FAQ entries, and business details are loaded into the agent's knowledge context.
- Webhook connection — the platform sets up webhooks so call data (recordings, transcripts, metadata) flows back to your dashboard in real time.
- Status update — the client's status changes to "Active (Agent Live)" and the agent appears on your Dashboard.
After Deployment
Deploying the agent is just the beginning. Here's what you can do to optimize the experience:
Fine-Tune the Agent
Go to the client's Voice Agent tab to adjust settings like the greeting message, personality, call handling rules, and more. See the Voice Agent Configuration guide for details.
Build the Knowledge Base
Upload documents to the Knowledge Base tab — service menus, pricing sheets, policies, or any other reference material. The more the agent knows, the better it handles nuanced questions.
Add FAQ Entries
If you know the most common questions callers will ask, add them as FAQ entries. This gives the agent direct, vetted answers for frequently asked questions.
Connect a Calendar
If your client takes appointments, connect a Cal.com calendar in the Connectors tab so the agent can check availability and book appointments directly during calls.
Tip: After deploying, make 3–5 test calls pretending to be different types of customers — one asking about hours, one requesting a service, one with an unusual question. This helps you identify any gaps in the agent's knowledge before real customers start calling.
Was this page helpful?