How a Phoenix HVAC company stopped losing after-hours calls and added $180K in service agreements in 90 days
The Problem
The company was losing 35–40% of after-hours service calls to competitors simply because no one answered. During summer peak season — when AC failures happen at 9pm on a Saturday — callers would hang up and dial the next company. A part-time answering service was costing $1,800/month and still missing calls. Their existing customer base was also largely untapped: thousands of past customers with aging equipment, no proactive outreach, and no service agreement program.
The Solution
We deployed an AI phone and SMS agent that handles every inbound call 24/7 — triages emergency vs. standard service, collects job details, schedules appointments directly into their dispatch system, and sends confirmation texts to both the customer and the on-call technician. In parallel, we built a proactive service agreement engine that analyzed their customer database by equipment age and service history, and launched an automated outreach campaign that converted past customers into recurring revenue.
The Results
- ↑After-hours lead capture went from ~60% to 97% — nearly zero missed calls
- ↑$180K in new service agreements signed in the first 90 days of proactive outreach
- ↑Average time from inbound call to scheduled appointment: under 4 minutes
- ↑Replaced $1,800/month answering service with a system that costs a fraction of that
- ↑On-call technicians get dispatched with full job context before they leave the driveway
The situation
Phoenix summers don't care about business hours. Air conditioners fail on Saturday nights. They fail on holiday weekends. They fail after the kids are already in bed and the house is 87 degrees.
This company had 11 experienced technicians, a strong reputation, and a real problem: when a customer called after 6pm, nobody answered. A part-time answering service was supposed to cover the gap, but missed calls were routine and the handoff to on-call techs was slow and inconsistent. By the time the callback came, the customer had already booked with whoever picked up first.
An estimated 35–40% of after-hours leads were walking to competitors. During a Phoenix summer, that's real money walking out the door.
The second problem was quieter but bigger. Thousands of past customers were in the system — homeowners with 8, 10, 12-year-old units — and there was no proactive outreach program. No service agreements. No maintenance reminders. Just waiting for the phone to ring when something broke.
What we built
24/7 AI phone and SMS agent
Every inbound call now reaches an AI agent built on the Claude API and connected to the phone system via Twilio. The agent handles the full intake conversation in a natural, unhurried tone — not an IVR menu, an actual conversation.
It determines whether the call is an emergency (no AC, no heat, water leak) or a standard service request, collects the relevant details about the system and the issue, checks technician availability in real time via the ServiceTitan integration, and books the appointment on the spot. A confirmation text goes to the customer immediately. A dispatch notification — with the customer's address, the problem summary, and relevant equipment history — goes to the on-call tech before they're off the couch.
For calls that need a human judgment call (complex commercial jobs, insurance-related work, customers who just want to talk to someone), the agent captures all the details and flags it for a same-day callback — with full context already pulled up.
The average time from call to confirmed appointment is under 4 minutes, around the clock.
Proactive service agreement engine
We pulled the customer database and built a scoring model based on three signals: equipment age, time since last service, and service history. Customers with units over 8 years old who hadn't been contacted in 12+ months scored highest.
The system launched a sequenced outreach campaign — a personalized email referencing their specific equipment, followed by an SMS two days later if no response. The message offered a fixed-price annual maintenance plan: one tune-up, priority scheduling, and a discount on any repairs.
Every response fed back into ServiceTitan automatically. Booked appointments appeared on the schedule with zero manual entry.
Technician dispatch with context
Previously, the on-call tech would get a text with an address and a vague description like "AC not working." Now they receive a structured dispatch message: the customer's name, address with a one-tap directions link, equipment make, model, and age pulled from service history, and a plain-English summary of what the customer described. They arrive knowing whether they need to bring refrigerant, a capacitor, or a float switch.
The field team noticed the difference within the first week.
The result
In the first peak season month after launch, after-hours lead capture went from roughly 60% to 97%. The part-time answering service was cancelled. On-call technician response time dropped significantly because the dispatch handoff was instant and the techs had what they needed before they left home.
The service agreement campaign converted a portion of the dormant customer base into recurring revenue: $180K in signed agreements in 90 days, with a pipeline of warm leads still being worked.
The owner still gets a morning briefing — a short daily summary of overnight calls, new appointments, and any flagged items that need attention. He reviews it with his coffee before the team starts their day.
"I was spending $1,800 a month on an answering service that still let calls slip through. Now every call gets handled the right way, and I've got a morning report waiting for me. It's the first time in 14 years this business runs the same whether I'm watching or not."
Want to see what this would look like for your business? Get your free AI Readiness Assessment or book a discovery call.