Multi-truck dispatch + bilingual AI receptionist for a US plumbing contractor

The client

A US plumbing contractor running 22 trucks across residential emergency service and scheduled commercial work in a major US Sunbelt metro. ~70% of the operation is residential emergency (water leaks, drain backups, water heater failures, sewer line emergencies). ~30% is scheduled commercial — drain maintenance contracts, backflow testing, fixture installs on commercial construction projects.

The owner had been on ServiceTitan for four years and had no plans to leave — the CRM was working, technician mobile worked, billing worked. The single problem driving revenue loss was the phone. Generic answering services couldn't qualify a 2 a.m. water-heater leak vs a scheduled drain cleaning request. The Spanish- speaking customer base — large in their market — was getting handled poorly when bilingual reception staff weren't available. After-hours emergency calls were going to voicemail roughly 6–10 times per week.

What the contractor needed

• 24/7 call coverage with trade-specific intake — distinguish 'water leaking now' from 'scheduled water heater replacement' from 'commercial backflow testing inquiry'.
• Bilingual English / Spanish — same dispatch logic, separate intake scripts, full call handling in either language.
• ServiceTitan integration — they did not want to migrate. The AI receptionist had to plug into the existing ServiceTitan tenant cleanly.
• Emergency triage — fire/flood/no-water/no-heat calls routed to on-call dispatch immediately; routine calls booked into the next-available scheduled slot.
• Service-area enforcement — automatically decline or route out-of-area calls (the operation serves specific zip codes; further-out calls get politely declined or referred).
• Parts-on-truck dispatch awareness — when assigning emergency calls, the system should know which trucks have which parts loaded, routing emergencies to trucks that can complete the job.

What Xenara built

An AI receptionist + dispatch automation overlay sitting on top of the existing ServiceTitan tenant. Built in 6 weeks. No migration. ServiceTitan continued to handle CRM, billing, technician mobile, and reporting. Xenara handled the phone + the dispatch logic.

• Bilingual AI receptionist — 24/7 call answering in English and Spanish. The system auto-detects language preference from first response and continues in that language. Intake script trained on 45 hours of recorded inbound calls (English and Spanish).
• Plumbing-specific intake — system asks the right qualifying questions: where is the leak, how long has it been leaking, is water shut off, what kind of water heater, is this commercial or residential, do you have a service contract with us. The output is a structured ticket ready for dispatch.
• Emergency routing logic — fire/flood/no-water emergencies trigger immediate on-call notification. Routine bookings drop into the next-available slot in ServiceTitan. Hold-for-callback queue for ambiguous calls.
• ServiceTitan API integration — every booked job creates an appointment in ServiceTitan with the right job type, customer record, address, and notes. Two-way sync — ServiceTitan job status changes flow back to the AI for follow-up customer SMS.
• Parts-on-truck dispatch overlay — emergency calls route to trucks carrying the right parts. The dispatcher sees AI suggestions: "Assign to Truck 6 — 12 minutes away + has the right water heater stock." The dispatcher still owns the decision.
• Service-area enforcement— auto-decline calls outside the operator's zip-code service map. Out-of-area calls get a polite explanation + referral to a partnered operator where available.
• Post-call SMS confirmation — every booked job gets an SMS with appointment time, technician name (when assigned), and a rescheduling link. Customer-facing live ETA when the technician hits "on the way" in ServiceTitan mobile.
• Call recording + transcription — every call recorded, transcribed, and indexed against the customer record in ServiceTitan. The owner reviews high-value transcripts weekly.
• Owner dashboard— answered / booked / declined / transferred metrics. Daily summary, weekly trend, monthly P&L tie-back. Specific reporting for Spanish-language calls vs English.

Estimated $190k/year in recoverable after-hours emergency revenue, calculated against the operator's pre-AI baseline: 6–10 missed emergency calls per week × avg $700 captured job value × 50 weeks × ~70% capture rate.

Results

• 22 trucks dispatched with parts-on-truck-aware AI suggestions; dispatcher retains override on every decision.
• 24/7 bilingual coverage — every after-hours call answered. Spanish-language calls now handled at parity with English, recovering customer segment that had been lost to competitors.
• Estimated $190k/year in recoverable after-hours emergency revenue.
• 4-week payback on the engagement — recovered revenue in the first month exceeded the one-time build cost.
• ServiceTitan workflow preserved — no migration, no retraining of office staff, no disruption to billing or technician mobile.
• Emergency routing under 90 seconds— from caller saying "the water is everywhere" to on-call technician's phone ringing.

Stack + delivery model

AI receptionist built on Anthropic Claude + custom voice layer with sub-1.2s first-response latency. Bilingual script tuned weekly from real call transcripts. ServiceTitan integration via official API + Marketplace. 6-week build from kick-off to live deployment. Soft launch on after-hours and overflow first; 24/7 cutover at week 5 after quality verification. Ongoing platform retainer covers script tuning, model updates, and seasonal demand handling.

Related

• AI Receptionist
• Field Service Management
• Dispatch Software
• Industries · Plumbing
• ServiceTitan Alternative