Every plumbing owner hits the same wall. The phone rings while you are under a house, in a crawl space, or elbow-deep in a water heater swap. You let it go to voicemail. The caller — whose basement is taking on water — does not leave a message. They call the next plumber on the list. By the time you surface, the job, and the customer, belong to somebody else.
So the question arrives, usually at the kitchen table at 10 p.m.: should I hire a receptionist, or use AI to answer the phones? Both answers are right — for different shops. This guide gives you the real cost math, the coverage problem nobody prices in, and a simple call-volume test so you can decide with your own numbers instead of a vendor's pitch. It also includes the honest section most AI sellers skip: the jobs a human dispatcher still does better, and the moments you should never hand to a machine.
Start with the real cost of the human option
The U.S. Bureau of Labor Statistics puts median pay for receptionists at $18.27 an hour — $38,010 a year (May 2025, Occupational Outlook Handbook). That is the middle of the national market. Dispatch experience, a major metro, or a tight labor market pushes it higher.
The wage is not the cost. The cost is the wage plus everything attached to it:
- Base wage: $38,010/yr at the BLS median
- Employer payroll taxes: Social Security and Medicare alone add 7.65% before state unemployment insurance and workers' comp
- Benefits and paid time off: even a lean package — some PTO, no health coverage — adds real money; a competitive one adds a lot
- Turnover: reception and dispatch seats turn over, and every departure restarts recruiting, training, and a stretch of phones answered by whoever is nearest
Add it up and a median-wage receptionist typically lands around $43,000–$48,000 a year all-in, roughly $3,600–$4,000 a month (illustrative range — your state, benefits, and comp rates move it). That is not an argument against hiring. A great dispatcher is worth every cent of it, and we will get to why. It is an argument for knowing the number before you compare it to anything else — because you pay that full amount in your slowest month exactly the same as your busiest.
The coverage problem: the math that actually decides this
Here is the figure that changes the decision for most plumbing shops. A full-time receptionist works about 40 hours a week. A week has 168 hours. One full-time hire covers roughly 24% of the week. The other 76% — every evening, every weekend, every holiday, every sick day and lunch break — the phone is exactly as uncovered as it is today.
Plumbing does not respect office hours. Burst pipes, sewer backups, and dead water heaters cluster in the hours nobody is staffed: early mornings, evenings, weekends, and the coldest night of the year. An AI answering setup, by contrast, is priced flat and answers at 2 p.m. and 2 a.m. for the same cost — AI reception and answering tools for small businesses commonly run in the $29–$300 a month range depending on features and call volume (typical market range at time of writing; check current pricing before you buy). Even at the top of that range, a full year of AI answering costs less than one month of a full-time hire.
The quotable version: A full-time receptionist covers 40 of the week's 168 hours — about 24%. The Bureau of Labor Statistics puts the median wage alone at $38,010 a year before taxes and benefits. Plumbing emergencies live in the other 76% of the week. The hire-vs-AI question is really a coverage question.
That does not settle it. It reframes it. You are not choosing between a person and a program. You are choosing how to cover 168 hours — and the honest answer for many shops is a mix, which we will get to after the test that tells you where you stand.
The call-volume test: three bands, three different answers
Count your inbound calls for two normal weeks — your phone system or carrier bill already has this number. Then find your band. (These bands are a decision framework based on the economics above, not industry statistics — run them against your own average ticket before treating them as rules.)
Band 1 — Under about 15 calls a week, mostly during work hours. You do not have a staffing problem yet; you have an interruption problem. A full-time hire is hard to justify at this volume — you would be paying $3,600+ a month (illustrative, from the math above) to answer three calls a day. AI answering for the overflow and after-hours, plus a tight callback routine for anything it cannot handle, covers this band for a small fraction of the cost.
Band 2 — Roughly 15 to 60 calls a week, with a real after-hours share. This is the hybrid band, and it is where most growing plumbing shops actually live. The calls are too many to absorb yourself and too judgment-heavy at peak times to hand over entirely. The pattern that works: a human — even part-time — owns the busy daytime window, scheduling judgment, and customer relationships; AI owns evenings, weekends, overflow when the line is busy, and the first capture on every emergency call. Each side covers the other's gap.
Band 3 — Over about 60 calls a week, multiple trucks, real dispatch complexity. Hire the human. At this volume you are not buying phone answering, you are buying a dispatcher: someone who knows your techs, your service area, your parts runs, and which customer gets priority when two emergencies land at once. AI becomes the dispatcher's backup — nights, weekends, and overflow — not the front line.
What AI can't do (the section AI sellers skip)
An AI answering tool is genuinely good at a narrow set of things: picking up instantly, every time; capturing a caller's name, address, problem, and urgency in a consistent format; answering routine questions (hours, service area, whether you do water heaters); and handing you a clean summary instead of a half-heard voicemail. For after-hours emergency capture — where the alternative is literally nobody — that is a large, real improvement.
Here is what it cannot do, and pretending otherwise is how shops get burned:
- It cannot judge a weird situation. A panicked caller describing a smell, a sound, and a puddle that do not add up needs a human who has heard a thousand calls. AI triages by script and keywords; it will miss the one call in fifty where the script is wrong.
- It cannot build a relationship. Your repeat customers — the landlord with twelve units, the restaurant that calls every quarter — stay loyal partly because a person knows them. Route your best accounts to a human voice, always.
- It cannot negotiate, soothe, or improvise. The furious customer with a flooded kitchen and a wedding tomorrow does not need a perfectly captured message. They need a person who can calm them down and move mountains on the schedule.
- It cannot dispatch. Knowing that your senior tech is twenty minutes from that address with the right part on the truck is operational judgment, not information capture.
- It can be wrong confidently. Any AI setup needs a human review loop: read the summaries every morning, correct the recurring mistakes, and keep a one-tap path for callers to reach a person. If a vendor tells you none of that is necessary, that is your answer about the vendor.
When the human hire is simply the right call
Hire the receptionist — proudly, and without treating AI as the "smart" alternative — if most of these are true: your call volume sits in Band 3; your average job value is high enough that one saved customer pays a month of payroll; your customers are repeat commercial accounts who expect a person; or you, the owner, are spending ten-plus hours a week on scheduling and callbacks that a good hire would take off your plate entirely. A dispatcher who also owns scheduling, review requests, and estimate follow-up is not a $38,010 expense. They are the person who lets you run a second truck.
A two-week test before you commit to either
Do not decide this from an article — including this one. Run this instead. Week one: log every inbound call with time, whether it was answered live, and what it was worth if it booked. Most owners who do this are surprised twice: by how many calls land outside business hours, and by how many daytime calls they "answered" while distracted and never followed up. Week two: price both options against that log — the all-in monthly cost of a hire (use the BLS median above as your anchor, adjusted for your market) versus an AI answering setup covering the hours your log says you actually miss. The option that covers your real gaps for the least money, without handing your best customers to a script, is your answer. For most small plumbing shops, the log points to the same place: AI on the 76% of the week nobody can staff, humans where judgment and relationships earn their pay.
Where to start
If your log says coverage is the problem, start by looking at tools built for exactly that gap — answering, capturing, organizing, and following up — before you commit to a full-time seat. Browse the AI Business Tools collection at JOJO WORLD for the current lineup, compare what each one actually covers against your two-week call log, and keep the human hire for the moment your volume — not your frustration — says it is time.