Skip to Content
DocumentationVoice AgentOutbound Calling

Outbound Calling (Calling Agent)

Your AI calling agent automatically calls new leads within 5 seconds of inquiry—increasing response time, boosting conversion rates, and booking appointments while leads are still engaged.

Why Outbound Calling Works

Speed matters: Leads contacted within 5 minutes are 21x more likely to convert than those contacted after 30 minutes.

Your AI calling agent:

  • ✅ Calls new leads in under 5 seconds
  • ✅ Works 24/7 (even at 2 AM)
  • ✅ Books appointments during the first call
  • ✅ Never forgets to follow up
  • ✅ Handles objections naturally

Result: 3-5x higher conversion rates compared to manual calling or delayed responses.


How It Works

Automatic Lead Calling Flow

  1. Lead comes in from Yelp, Thumbtack, website form, etc.
  2. AI calls immediately (within 5 seconds)
  3. Personalized intro - References their specific inquiry
  4. Qualifies the lead - Confirms service need
  5. Answers questions - Quotes pricing from knowledge base
  6. Books appointment - Checks calendar and schedules
  7. Logs everything - Full transcript, recording, and extracted data

Average call duration: 2-4 minutes
Connection rate: 65-75% (depends on time of day)
Booking rate: 40-50% of connected calls result in appointments


Setting Up Outbound Calling

Create a Calling Agent

  1. Go to Voice Agent → Create New Agent
  2. Select “Calling Agent (Outbound Calls)”
  3. Configure basic details:
    • Company name
    • Agent name
    • Voice selection
    • Main goal (what the AI tries to achieve)

Configure Call Context

Your AI needs to know why it’s calling each lead. You’ll provide:

Lead variables (automatically populated):

  • {{lead_name}} - Lead’s name
  • {{lead_phone_number}} - Lead’s phone
  • {{service_title}} - Service they requested (e.g., “Garage Door Repair”)
  • {{service_description}} - Details about their request

Example lead context:

- Lead name: John Smith - Phone: +1-909-555-1234 - Service title: AC Repair - Service description: Customer's AC stopped working during heatwave, needs urgent repair

Set Your Main Goal

The “Main Goal” tells the AI what to accomplish on the call.

Example goals:

You're calling because the customer requested AC repair service on Yelp. Confirm their service need, answer any questions, and schedule an appointment.
You're following up on the customer's garage door repair inquiry. Quote pricing if asked, and book an appointment as soon as possible.

Best practice: Keep the main goal under 50 words and focused on the outcome (usually booking an appointment).

Configure Services & Pricing

Same as answering agent:

  • Enable service identification
  • Add services to knowledge base
  • Upload pricing information
  • Configure “Free on-site estimates” if applicable

See Inbound Call Handling for service configuration details →

Set Up Appointment Scheduling

Connect your calendar system:

  • Cal.com, FieldPulse, or GoHighLevel
  • Provide API keys
  • Configure collection fields (name, phone, address)

Connect to Lead Sources

Enable automatic calling from:

NZ Leads Integration (if you use NZ Leads for text-based lead management):

  • New leads from Yelp, Thumbtack, etc. are automatically called
  • Toggle: “Auto-call new leads” in Voice Agent settings

Zapier/Webhooks (for custom integrations):

  • Trigger AI calls from any platform
  • Send lead data via webhook
  • [Contact support for webhook documentation]

Test with Manual Call

Before going live, make a test call:

  1. Go to Voice Agent → Calling Agents
  2. Click “Make Test Call”
  3. Enter a phone number and lead context
  4. Review call log after the call

Call Flow Breakdown

1. Opening & Introduction

The AI calls the lead and introduces itself with context:

Example opening:

“Hi, this is Sarah from ABC Plumbing calling about your service request. You mentioned your water heater stopped working - how can I help?”

What makes this effective:

  • ✅ Mentions business name immediately
  • ✅ References their specific inquiry (“water heater stopped working”)
  • ✅ Opens with helpful tone, not salesy

Handling voicemail: If the call goes to voicemail:

  • AI detects answering machine
  • Waits for beep
  • Leaves configured voicemail message:

    “Hi, this is Sarah from ABC Plumbing returning your call about water heater service. Please call us back at 909-555-1000. Thanks!“


2. Confirming Service Need

The AI references the lead’s original request:

AI:

“You requested garage door repair—how can I help?”

What the AI does:

  • Uses {{service_title}} from lead context
  • Asks open-ended question to let customer explain
  • Does NOT re-ask “What service do you need?”—they already told you!

3. Qualifying & Answering Questions

The AI handles common questions:

Pricing:

  • Quotes from knowledge base
  • Asks clarifying questions if needed (“Single or double door?”)
  • Offers free estimate if pricing unavailable

Availability:

  • “When can you come out?”
  • AI checks calendar and offers specific times

Service area:

  • “Do you service my area?”
  • AI answers from knowledge base (if configured)

Urgency:

  • “This is an emergency!”
  • AI prioritizes and marks appointment as urgent

4. Appointment Booking

Same as inbound calls:

  • Collects necessary info (name, phone, address)
  • Checks calendar availability
  • Offers 2-3 time slots
  • Books when customer agrees
  • Confirms appointment details

See Inbound Call Handling for detailed booking flow →


5. Handling Objections

The AI is programmed to handle common objections:

“I’m just getting quotes”

“That’s completely understandable. I can provide a quote over the phone, or we offer free on-site estimates. Which works better for you?”

“Can someone call me back later?”

“Of course! What’s the best number to reach you, and is there a preferred time?”
→ AI takes note and logs for callback

“I already found someone”

“Glad you found help! If anything changes, feel free to reach out.”
→ AI ends call politely

“How much will this cost?”

[Quotes from knowledge base]
→ Then: “Would you like to schedule an appointment?”


Voicemail Strategy

Voicemail Message Configuration

Customize what the AI says when it reaches voicemail:

Default message:

Hi, this is [AGENT_NAME] from [COMPANY_NAME] returning your call about [SERVICE]. Please call us back at [BUSINESS_PHONE]. Thanks!

Custom example:

Hi, this is Sarah from ABC Plumbing. I'm calling about your garage door repair request. We have same-day availability—give us a call back at 909-555-1000 to schedule. Thanks!

Best practice: Keep voicemails under 15 seconds. Include a clear callback number and mention urgency (“same-day availability”) if relevant.


Timing & Best Practices

When to Call

Optimal calling times:

  • ✅ Weekdays: 10 AM - 12 PM, 2 PM - 5 PM
  • ✅ Weekends: 10 AM - 3 PM
  • ❌ Avoid: Early mornings (before 9 AM), late evenings (after 7 PM)

Immediate calling (within 5 seconds):

  • Best for leads from Yelp, Thumbtack, Google LSA
  • Leads expect a quick response
  • High conversion rates

Delayed calling:

  • For website forms submitted outside business hours
  • AI can call the next morning
  • Still faster than manual follow-up

Call Pacing

If you have multiple leads in queue:

  • AI paces calls to avoid overwhelming you
  • Configurable: 1-5 calls per minute
  • Ensures appointments don’t overlap

Retry Logic

If the call doesn’t connect:

  • 1st attempt: Immediate (within 5 seconds)
  • 2nd attempt: 15 minutes later
  • 3rd attempt: 2 hours later
  • Final attempt: Next business day

Voicemail strategy:

  • Leave voicemail on 1st and 3rd attempt only
  • Don’t spam with multiple voicemails

Integration with NZ Leads

If you use NZ Leads for text-based lead responses:

Automatic Workflow

  1. New lead arrives from Yelp/Thumbtack
  2. NZ Leads sends welcome text (if configured)
  3. AI calls the lead within 5 seconds
  4. Lead is marked in dashboard with call status
  5. You get notification if appointment is booked

Unified Dashboard

View all leads in one place:

  • Text conversations - See messages from NZ Leads
  • Call logs - See AI call transcripts and recordings
  • Appointments booked - Both text and call bookings
  • Lead status - Track which leads have been called

Handoff Between Text and Calls

  • Lead responds to text but doesn’t book → AI calls
  • Lead doesn’t answer call → NZ Leads continues text follow-up
  • Seamless experience for the customer

Batch Import & Manual Calling

Manual Single Call

To call a specific lead:

  1. Go to Voice Agent → Make Call
  2. Enter:
    • Phone number
    • Lead name
    • Service title
    • Service description
  3. Click “Call Now”
  4. AI calls immediately

Bulk CSV Upload

For calling multiple leads at once (e.g., old customers):

See Batch Campaign Calling documentation


Call Logs & Analytics

After each call, view:

Call Details:

  • Full transcript
  • Audio recording
  • Call duration
  • Connection status (answered, voicemail, no answer)
  • Outcome (appointment booked, callback requested, not interested)

Lead Data Extracted:

  • Name
  • Phone
  • Service requested
  • Appointment details
  • Customer sentiment (positive, neutral, negative)

Learn more about Call Logs & Analytics →


Pricing

Outbound Calling Costs:

  • Per-minute rate: $1/minute for all calls
  • Phone number (if needed): $10/month

Example monthly cost:

  • 100 calls per month
  • Average 3 minutes per call
  • = 300 minutes × $1/min = $300 in call costs
  • Phone number: $10/month
  • Total: $310/month

ROI calculation:

  • If AI books 20 appointments/month
  • Average customer value: $300
  • Revenue: 20 × $300 = $6,000
  • Cost: $310
  • ROI: 19x return

Advanced Features

Custom Calling Hours

Set when the AI should call:

  • Business hours: Call immediately
  • After hours: Wait until next morning
  • Weekends: Call or wait until Monday

Priority Calling

Mark certain lead sources as high-priority:

  • Emergency service requests → Call immediately
  • Google LSA leads → Call within 30 seconds
  • Website forms → Call next business day

Call Recording & Compliance

All calls are recorded by default:

  • Stored for 90 days
  • Downloadable as MP3
  • Includes AI disclosure at start of call (where required by law)

A/B Testing

Test different approaches:

  • Agent voice: Male vs. Female
  • Opening script: Reference service vs. generic greeting
  • Main goal: Book appointment vs. qualify lead first

Troubleshooting

AI Isn’t Calling New Leads

Check:

  • ✅ Calling agent status is “Active”
  • ✅ “Auto-call new leads” toggle is ON
  • ✅ Lead source integration is connected
  • ✅ Lead has valid phone number

Low Connection Rate

Improve connection rates:

  • Call during optimal hours (10 AM - 5 PM)
  • Enable retry logic (2-3 attempts)
  • Leave compelling voicemails
  • Reduce time between lead submission and call (under 1 minute is best)

Low Booking Rate

Optimize for more bookings:

  • Review call transcripts - see where AI loses leads
  • Add common objections to knowledge base
  • Enable pricing (transparency helps)
  • Offer flexible appointment times
  • Train AI on specific service benefits

Best Practices

✅ Do This

  • Call new leads within 5 seconds (speed matters)
  • Reference their specific inquiry in the opening
  • Leave voicemails on 1st and 3rd attempt only
  • Review call logs weekly to optimize
  • A/B test different greetings and voices
  • Enable calendar integration for instant booking

❌ Avoid This

  • Don’t call outside business hours (unless emergency services)
  • Don’t leave voicemail on every attempt (spam = bad)
  • Don’t make the opening too salesy
  • Don’t re-ask questions the lead already answered
  • Don’t call leads who requested “text only”

Real-World Example

Scenario: HVAC company gets Yelp lead at 2:34 PM on a Tuesday

What happens:

  1. 2:34:05 PM - AI calls the lead
  2. 2:34:15 PM - Lead answers: “Hello?”
  3. AI: “Hi, this is Sarah from Cool Air HVAC calling about your AC repair request. You mentioned your AC stopped working—how can I help?”
  4. Lead: “Yeah, it’s really hot in here. How much will a repair cost?”
  5. AI: [Checks knowledge base] “An AC service call is $89, which includes diagnostics. If repair is needed, parts and labor depend on the issue. We offer same-day service—would you like to schedule?”
  6. Lead: “Yes, today if possible.”
  7. AI: [Checks calendar] “I have a tech available between 4 and 6 PM today. Does that work?”
  8. Lead: “Perfect.”
  9. AI: [Books appointment] “You’re all set! A technician will be at 123 Main Street today between 4 and 6 PM. You’ll get a text confirmation shortly.”
  10. 2:37 PM - Call ends. Total duration: 2 min 45 sec.

Result: Appointment booked 3 minutes after lead came in.


Next Steps

  1. Create your calling agent - Getting Started Guide →
  2. Configure services - Agent Configuration →
  3. Add pricing to knowledge base - Knowledge Base Guide →
  4. Enable auto-calling - Connect to NZ Leads or Zapier
  5. Monitor performance - Call Logs & Analytics →

Questions? Contact support at support@nzleads.com or use the in-app chat.

Last updated on