Services Tools Industries Case Studies Pricing NewsBlogAbout Book Free Audit
HVAC / Home Services Tucson, AZ, USA Deployed

// Case Study

A full-stack AI lead engine: two voice agents, an SMS engine, and a 5-system booking chain.

Six interconnected systems turn inbound calls, web-form leads, and a 4,000-contact database into confirmed ServiceTitan bookings with zero dispatcher involvement, from form fill to booked job.

0
Systems built
24/7
Call coverage
0+
Contacts reactivated
<10s
Form fill to booking
Goodfellas AC / Lead engine Live in production
Inbound call Web-form lead 4,000 database n8n + Retell AIqualify + route ServiceTitan

The problem

Leads were going to whoever called first.

Goodfellas was losing HVAC leads to competitors who called faster. Web-form submissions sat uncontacted for hours, peak-season inbound calls went unanswered, and thousands of dormant contacts were never re-engaged.

Every qualification call needed a dispatcher to walk five data points manually before ServiceTitan. After-hours calls hit voicemail and were lost, and no call analytics existed to see where leads dropped off.

What we built

Six interconnected systems on self-hosted n8n, orchestrating Retell AI voice, GoHighLevel, and ServiceTitan.

01

Outbound Kate

A GHL form triggers an n8n flow that normalises the phone number to E.164 and fires a Retell AI call within minutes, using a 5-question qualification script with 4-bucket routing.

02

5-system booking chain

Retell to n8n to GHL (find or create contact, native booking) to a Zapier webhook to ServiceTitan, completing in under 10 seconds with a fallback booking link.

03

Inbound Kate

The same qualification and routing, running 24/7, with an after-hours flag that changes both the greeting and the routing logic automatically.

04

Post-call analytics

Every call logged to PostgreSQL with 15+ fields (duration, recording, sentiment, routing category, issue and unit type, urgency, booking status) with a 7-branch GHL follow-up.

05

Text engine

Four n8n workflows pulling 60-message history from PostgreSQL, routing to 4 AI agent types with 6 booking functions, backed by a knowledge base of vector embeddings in Supabase.

06

Database reactivation

A CSV upload triggers a multi-touch per-lead sequence (SMS, AI call, SMS, email), with any reply immediately exiting the contact from the sequence.

Results

Zero dispatcher time on lead qualification.

Inbound and outbound calls handled 24/7 with automatic after-hours detection.

4,000+ dormant contacts cycled through a structured multi-touch sequence.

A 5-system booking chain completing in under 10 seconds with a fallback booking link.

15+ fields of call intelligence logged per call.

A text engine routing to 4 specialist agents with 60-message memory and live booking functions.

Built with

GoHighLevelGoHighLevel n8nn8n Retell AIRetell AI ZapierZapier ServiceTitanServiceTitan PostgreSQLPostgreSQL SupabaseSupabase OpenAIOpenAI TwilioTwilio

Related case studies

// Free Audit

Want a lead engine like this in your business?

Book the audit. Thirty minutes. We map what your business can automate.

No obligation Reply within 24 hours Fixed-price or hourly