Build a Plumbing AI Follow-Up System You Can Sell on Retainer

Introduction

Plumbing companies generate a lot of leads and lose a lot of them to slow follow-up. A system that turns dead lead lists into booked jobs automatically is easy to demonstrate, easy to price, and easy to position as a monthly retainer service.

In this guide, we build a fully automated plumbing lead follow-up system using Vapi (AI voice agents) and n8n (workflow automation). The system goes through old plumbing leads every morning, calls them by name, qualifies whether they still need help, checks real calendar availability, books appointments automatically, and sends a confirmation text — without anyone on the team touching it.

Video Tutorial

Watch the full build with a live demo of the system calling and booking:

What This System Does

The system runs on two platforms working together:

• Vapi — handles the AI voice agent that makes outbound calls, has natural conversations, and triggers tool calls
• n8n — handles the backend workflows: pulling leads from Google Sheets, routing tool calls, checking calendar availability, booking appointments, and sending SMS confirmations

Every morning, the n8n outbound dialer pulls leads from a Google Sheet, loops through each one, and triggers Vapi to call them. The AI agent introduces itself, asks if they still need plumbing help, qualifies the issue, and if they do — checks the calendar and books an appointment right on the call.

Setting Up the Vapi Assistant

Create a new Vapi assistant configured for outbound calling. Key settings:

• Voice: Choose a natural-sounding voice (not robotic) — this is the first impression
• Voicemail detection: Enable so the system skips voicemails and moves to the next lead
• Call forwarding: Configure a fallback if the customer asks to speak to a human

Plumbing System Prompt Breakdown

The system prompt is the core of the AI agent's behavior. For a plumbing follow-up agent, it needs to:

• Introduce itself as calling from the plumbing company (by name)
• Reference the customer by name and their original inquiry
• Ask if they still need plumbing help
• Qualify the issue type (leak, clog, drain backup, water heater, etc.)
• Determine urgency
• Offer to book an appointment if they're interested
• Handle common objections naturally

The 3 Vapi Tools

The voice agent has three tools it can call during the conversation:

1. Client Lookup — pulls the customer's details from Google Sheets so the agent can reference their original issue
2. Check Availability — queries Google Calendar for open appointment slots
3. Book Appointment — creates the calendar event and triggers the SMS confirmation

n8n Backend Setup

The n8n backend consists of two main workflows:

Outbound Dialer Workflow

This workflow runs on a schedule (every morning at 9 AM). It pulls the lead list from Google Sheets, filters for leads that haven't been contacted yet (or haven't been reached), and triggers Vapi to call each one sequentially. After each call, it updates the Sheet with the call status.

Tool Call Webhook Workflow

When the Vapi agent needs to look up a client, check availability, or book an appointment, it sends a tool call to n8n via webhook. The webhook workflow routes the request to the right subworkflow:

• Client lookup subworkflow: Searches Google Sheets by phone number and returns client details
• Availability subworkflow: Checks Google Calendar for open slots in the next few days
• Booking subworkflow: Creates a Google Calendar event and sends an SMS confirmation via Twilio

Google Sheets & Calendar Integration

Google Sheets acts as the lead database — each row has the customer's name, phone number, original issue, date of inquiry, and contact status. Google Calendar is the booking system — the AI checks for open slots and creates events with the customer's details.

One practical issue covered in the video: daylight savings time bugs. If your timezone handling isn't explicit, you'll get off-by-one-hour booking errors. Always use explicit timezone offsets in your n8n datetime nodes.

Why This Offer Works for Plumbing Companies

Plumbing leads are time-sensitive. When someone's pipe bursts or their water heater dies, they call multiple companies and book whoever responds first. But plumbing companies also accumulate large lists of leads who called, didn't book, and were never followed up with.

This system turns those dead lead lists into booked jobs automatically. The ROI story is simple: "You have 200 old leads sitting in a spreadsheet. This system calls all of them this week. If it books even 5 jobs, that's $2,000+ in revenue from leads you already paid for."

Frequently Asked Questions

How much does it cost to run this system?

Vapi charges per minute of call time (typically $0.05–0.10/min). n8n is free if self-hosted. Twilio SMS is ~$0.0079/message. For a batch of 50 calls, you're looking at $5–10 in total costs. If one of those calls books a $400 plumbing job, the ROI is massive.

What if the customer picks up and doesn't want to talk?

The AI agent is trained to be polite and brief. If the customer says they're not interested, it thanks them and ends the call. The Sheet is updated so they're not called again.

Can I use this for other service businesses?

Absolutely. The architecture is the same for HVAC, electrical, roofing, landscaping, auto repair — any service business with a lead list and an appointment-based workflow. Swap the system prompt and you have a new vertical.

How do I pitch this to a plumbing company?

Don't say "AI." Say: "I have a system that calls your old leads every morning, qualifies them, and books appointments on your calendar automatically. You just show up." Demo it live on a test lead. The response rate from the demo alone usually closes the deal.