How to Set Up and Optimize Voice AI Agents

Overview

Voice AI Agents can handle inbound calls for your business and take actions based on the goals you set—collecting information, routing calls, triggering workflows, sending notifications, and more. With the right configuration, your agent can support customers 24/7, reduce missed calls, and improve speed-to-lead without adding workload to your team.

Key Features & Benefits

• Automated call handling: Answer inbound calls and respond using prompts and goals you configure.
• Higher call throughput: Voice AI supports up to 200 calls per minute for simultaneous interactions (regular call limits remain unchanged).
• Flexible goals: Use Basic Mode to collect core details, or Advanced Mode for deeper logic and actions.
• Workflow integration: Trigger workflows, send SMS, or update fields based on call outcomes.
• Voice and tone customization: Choose voices/accents to match your brand and improve caller experience.
• Availability controls: Assign phone numbers, set working hours, and use AI as backup when humans don’t answer.

Step 1: Access the AI Agents Section

1. Navigate to Settings → AI Agents.
2. Click + Create Agent to start.

Step 2: Configure Agent Details

Set the basics for how the agent answers calls, what voice it uses, and how it greets callers.

• Agent Name: Use a unique name (example: “Customer Support Bot”).
• Business Name: Confirm the business name used in greetings.
• Language: Select the language for call handling.
• Voice Selection: Choose and preview voices with the play button.
• Agent Direction: Typically Inbound.
• Initial Greeting: Example: “Hello, you’ve reached [Business Name]. How can I help you today?”

Step 3: Define Agent Goals (Basic vs Advanced)

Actions You Can Configure

Actions help your agent do real work during the call and after the call ends.

• Call Transfer: Transfer to a human based on conditions.
• Workflow Triggers: Start workflows based on call outcomes.
• Update Contact Fields: Save collected data into the contact record.
• Send SMS: Send a message during or after the call for follow-up.
• Appointment Booking: Book appointments automatically (calendar rules and slot logic apply).
• Custom Actions (Webhooks): Trigger external systems with POST webhooks (supports auth/headers/dynamic parameters).

Step 4: Configure Phone & Availability

• Assign phone numbers: Select the numbers that should route to this agent.
• Backup handling: Enable AI as backup if the user or forwarding number does not respond.
• Number Pool support: Assign pools so calls can be routed based on availability and pool logic.
• Working hours: Define when the agent answers calls (and when humans should take calls).

Step 5: Test and Finalize

1. Place a test call to confirm the agent behaves as expected.
2. Review Call History, Transcripts & Recordings, and any summaries/feedback.
3. Open call logs to confirm:
   • Caller phone number (and contact name if saved)
   • Triggered actions (workflow, SMS, transfer, booking)

Step 6: Use Analytics to Optimize Performance

Use analytics to measure what’s working and adjust prompts, goals, or actions for better outcomes.

• Call analytics: Total calls, trends, and average call duration.
• Actions triggered: Breakdown by type and frequency.
• Updated contact fields: Most frequently updated fields during calls.
• Sentiment insights: AI-analyzed sentiment metrics (if enabled/available).
• Filters & comparisons: Date ranges, agent filters, and period-over-period changes.

Pro Tips

• Train your agent: Use clear prompts and examples for complex questions to improve accuracy.
• Use Advanced Mode for scenarios: Transfers, workflows, and custom actions help handle specialized call flows.
• Set working hours strategically: Let humans take calls during staffed hours and use AI as backup outside those hours.