How a 3-truck plumbing shop cut missed calls 70% in 14 days

The shop runs three trucks out of a single bay in central Texas — two techs in the field most days, a third on a swing rotation for new construction rough-ins. The owner-dispatcher handles the office: phones, scheduling, parts runs, and the email that always lands at lunch. The shop was busy when we started the engagement, but the call log told a different story than the trucks. Roughly seventy percent of calls that landed outside the office window were going to voicemail. The dispatcher would call them back at 7 a.m. on the next business day; about a third of those numbers never picked up.

A missed call for a residential plumbing shop is not theoretical loss. A burst pipe at 11 p.m. is the highest-margin job of the week — anyone who reaches voicemail and dials the next name on the Google results is gone. By the time the dispatcher called back the next morning, the homeowner had already found a competitor who picked up. Two or three of those per week compounded into a real revenue gap that the owner could measure against the truck schedule.

We did not start by replacing the phone system or ripping out Jobber. The shop already had a Jobber account, a Twilio number for the office line, and a Google Business listing feeding a steady fifteen to twenty inbound calls a day. The constraint was the gap from 5 p.m. to 8 a.m. weekdays plus weekends — and a dispatcher who could not pick up the office phone while she was elbow-deep in a service window. We mapped the missed-call points first: after-hours emergencies, weekend rough-in coordination calls, and the lead flow that came through the website form during business hours when the dispatcher was on a service window.

The first workflow we turned on was after-hours AI call answering. When the office line rang outside the dispatcher’s working hours, an AI agent picked up before the second ring, classified the request, and captured the structured intake a plumbing dispatcher needs: address, problem type (leak, clog, water heater, gas), property type, urgency, decision-maker on site, and callback preference. The agent booked the on-call tech into Jobber directly when the request was a true emergency, and wrote everything else as a structured lead so the dispatcher woke up to a queue she could work through on her own terms instead of an unfiltered voicemail box.

Inside the first week the dispatcher noticed the difference at 7 a.m. instead of 8:30. Week two, the on-call tech reported that the structured intake notes were saving him between ten and fifteen minutes per Saturday job — he was rolling onto the property already knowing the issue, the system age, and whether the homeowner was the decision-maker. The dispatcher estimated that call-backs she used to do in three days now fit into a single morning, because the AI was collecting the basics the tech would have asked for anyway.

We added lead capture from the website form two weeks in. Free-estimate requests from the site now land as a structured intake in Jobber with the same fields the after-hours calls produce: square footage, property age, decision-maker, urgency, preferred contact window. The leads that used to sit in a shared inbox as raw emails are now in the dispatcher’s queue with context attached, sorted by the urgency field the AI extracts. She stops spending the first hour of the day parsing email and starts spending it booking the work.

The metric the owner asked us to track was missed-call rate. The baseline measurement, taken from the Twilio log over the month before the engagement, was roughly seventy percent of after-hours calls going to voicemail with no callback answered within twenty-four hours. Fourteen days after turning on the AI workflow the same metric, measured the same way, sat at twenty-one percent. Of that residual, two thirds were callers who explicitly declined to leave a callback number and asked to call back during business hours — a different problem, not a missed-call problem. The owner’s conclusion: the after-hours funnel stopped leaking, and the calls that did leak were the calls that were never going to book anyway.

A side effect we did not anticipate: the dispatcher’s job got better. The week before, she had been triaging voicemail against an unanswered inbox and fielding a rear-view mirror checklist at every morning standup. Two weeks in, she was running the bookings queue with the same time budget and finding an hour a day to call back warm leads that the office had been too buried to surface. She has also, in her words, stopped answering the phone in a panic. The two-line triage that the AI does at the front of the call gives her the breathing room to make a thoughtful decision about whether a job is a fit for the shop.

The shop is not unique. Three-truck residential plumbing, three-truck HVAC, two-truck electrical — the shape of the answer is the same. The stack is whatever the shop already pays for (Jobber, ServiceTitan, Housecall Pro) plus a Twilio number they already use for the office line. The AI sits in front of the after-hours gap and writes structured intake into the system of record. The reads of the operation improves because the dispatcher stops being the phone tag bottleneck; the writes are faster because the AI has already asked the questions the tech would have asked anyway.

If your service-trades operation is running into the same pattern — busy trucks, quiet call log, dispatcher buried under morning voicemail — the right first step is to measure the after-hours gap honestly for thirty days. Twilio logs it, Jobber will tell you what percentage of after-hours calls resulted in a booked job, and the dispatcher’s calendar tells you the rest. From there the answer is small, not transformative: an after-hours intake workflow, a structured lead capture from the site, and a CRM write-back that lets the office see what the AI collected. We will walk a thirty-minute version of that conversation through your shop on a walkthrough call.