This is the delivery page rather than the research page. Three real plumbing situations the agent has to handle, the four-phase timeline from kickoff to live traffic, and the three numbers we commit to: 10 use cases mapped, $1,400–$4,800/mo savings range and 21-day implementation.
Three situations this has to handle
The Sunday-night burst pipe — 10:47pm Sunday in February. A pipe behind the laundry-room wall lets go. Water everywhere. Homeowner Googles 'emergency plumber near me' and your name is in the local pack. They call. The AI agent picks up in two rings, walks them through shutting the main valve while it qualifies (which fixture, how much water, is the breaker box flooding), then dispatches the on-call tech with a PEX kit. The tech is rolling by 11:02pm. Total damage is contained at maybe $4,000 of restoration. If the call had gone to voicemail, the homeowner would have called the next two plumbers in the pack — and your shop would have lost a customer who, statistically, is good for another $8,000 of plumbing service over the next five years.
The summer drain Saturday — Saturday at 11am. Twelve drain calls hit the line in an hour. Six are clear emergencies; six can wait until Monday. Your CSRs would have triaged maybe four of them and let the rest go to voicemail. The agent triages all twelve in parallel. The six real emergencies go on a truck within the hour. The six non-urgent calls get booked into Monday's calendar with a confirmation text. By Monday morning, your office team comes in to a fully booked week instead of a 48-message voicemail pile.
The annual membership push — You sold 940 annual plumbing-maintenance plans over the last three years. Two hundred ten of them are coming up for renewal this month. In the old workflow, your office team would have gotten through maybe 60. The outbound agent dials all 210 across two weeks, catches 130 live, and books spring drain treatments in the same conversation. Renewal capture for the month jumps from 35% to 70%, which translates to $25K+ of recurring revenue you would otherwise have lost to inattention.
The four phases
Week 1 — Discovery + scope. Map call types (drain, water heater, repipe, emergency), dispatch rules, on-call rotation, the price book, and the integrations. Document your top 20 dispatcher decisions.
Weeks 2–3 — Build voice agent against your dispatch system. Load shop-specific playbooks. Pilot internally on a forwarded test number; senior CSR roleplays 30+ calls.
Week 4 — Live pilot on overflow + after-hours. Senior CSR reviews recordings daily, tunes rules. Onboard memberships outbound + reviews automation.
Months 2–3 — Full rollout. Add the water-heater photo-quote agent, the dispatch-confirmation flow, and the sales coaching module. Weekly KPI review.
What the plumbing dispatcher opens with
"Hi, I have water coming out from under my kitchen sink and I don't know what to do." That is the first line, and it is deliberately a question rather than a greeting — the fastest way to find out whether this is an emergency or a price enquiry is to ask.
The three KPIs we commit to
10 use cases mapped
$1,400–$4,800/mo savings range
21-day implementation
Related trades we build for
Operators in hvac, electrical and roofing run a similar build, because the failure pattern is the same even when the vocabulary is not.