How a Plumber Grew Bookings 38% in 90 Days

Editor’s note: This case study is a composite, drawn from patterns we’ve seen repeatedly across plumbing companies in the $4M–$12M revenue range. It is illustrative — a representative example, not a single named customer. The numbers, however, reflect outcomes consistent with what mid-sized plumbing companies actually experience after implementing AI voice agents and demographic intelligence.

The Company

Call them “Tideline Plumbing.” Family-owned. Twenty-three years in business. Operating in a mid-sized metro on the East Coast with a population around 850,000. Seven trucks, eleven technicians, two service advisors, one office manager who’s been with the family since the beginning, and one owner — we’ll call him Mark — who knew, going into the year, that something had to change.

Tideline did roughly $6.8M in revenue the year before. Solid book of recurring service plan customers, good reviews, an established brand. By every external metric, a healthy plumbing company.

Internally, Mark could feel the ceiling.

Calls volume was up year over year — local SEO had been compounding, the company had run a few decent radio campaigns, and a long-running referral program kept the inbound steady. But booked jobs weren’t keeping pace. The pipeline was getting fatter without the revenue scaling proportionally.

Mark suspected the problem was on the phone. He was right. He just didn’t know how right.

The Baseline (Before Anything Changed)

Tideline gave us 90 days of call logs, dispatch records, and CRM exports to audit. Here’s what the data said.

Inbound call volume (rolling 90-day baseline):

• 2,847 total inbound calls
• 1,931 calls during business hours (M–F, 7 a.m.–6 p.m.)
• 916 calls outside business hours or weekends

Answer rate:

• Business hours: 84% answered
• After hours: 12% answered (a third-party answering service caught the rest, but only took messages)

Booking rate from answered calls:

• 41% of business-hours calls converted to a booked job
• 9% of after-hours messages eventually converted (most never called back, or called a competitor in the meantime)

Estimated revenue captured vs. left on the table:

• 90-day booked revenue: $1.71M
• Estimated 90-day lost-call revenue (using their average ticket of $612 and assuming a conservative 35% would have converted if answered and qualified): roughly $312,000 in 90 days, or $1.25M annualized

Mark stared at the $1.25M number for a long time.

“That can’t be right,” he said.

It was right. And the audit hadn’t even gotten to the calls his team had answered but had handled poorly.

The Diagnosis

We sat down with Mark, his office manager (we’ll call her Janelle), and his two service advisors and went through the four most common patterns we saw in the call audit.

Pattern 1: After-hours bleed.

Tideline’s answering service was, generously, a glorified voicemail. It took the caller’s name and number, sent Janelle an email, and Janelle would call back in the morning. By morning, 60–70% of those callers had already booked with someone else. For plumbing — where leaks, clogs, and water-heater failures are time-sensitive — this was catastrophic.

Pattern 2: Mid-day overflow.

Business-hours answer rate looked acceptable at 84% — until you broke it down by time of day. Between 9 a.m. and 11 a.m. on weekdays, the answer rate dropped to 67%. That’s the window where homeowners call after the morning rush. Janelle and the two advisors were physically incapable of answering everyone. Roughly 40 calls a week were lost in that two-hour window alone.

Pattern 3: Bad triage.

The team treated every caller the same. The $98 drain-snake caller and the $7,800 water-main replacement caller got the same 6-minute intake. Predictably, the high-value caller — who was already comparison-shopping — often hung up before they reached the scheduling step.

Pattern 4: No follow-up engine.

Estimates given but not booked within 24 hours had a 14% close rate. Industry benchmarks for plumbing with a real follow-up system are closer to 35–45%. There was no follow-up system. Janelle did her best on Fridays, but Fridays were chaos.

Four leaks. The math on closing each of them was hard to argue with.

The Plan

We mapped a 90-day implementation across three phases.

Phase 1: Days 1–14 — Foundation

• Port Tideline’s main number to Caller Technologies’ VoIP system
• Stand up the AI voice agent in parallel with the existing team — it would handle after-hours and overflow only at first, so the team could observe and trust it
• Connect caller intelligence: every inbound call now arrives at the agent (and the team) with the demographic profile pre-populated — address, property type, home value range, homeowner vs. renter, income range, distance to nearest crew, prior call history
• Build the smart routing rules: emergency keywords route directly to dispatch; high-value project keywords route to the senior advisor; out-of-area calls route to graceful referral; recurring plan customers get personalized greeting

Phase 2: Days 15–45 — Optimization

• Train the AI voice agent on Tideline’s actual service catalog, pricing tiers, scheduling windows, and emergency protocols
• Roll the AI voice agent forward into business hours: it handles initial triage on every call, and either books directly or routes to the human advisor based on call type
• Turn on call analytics dashboards for Mark and Janelle — daily reporting on answer rate, booking rate by call type, and revenue captured by demographic segment
• Set up AI coaching summaries: every human-handled call gets an automatic summary, sentiment analysis, and missed-opportunity flags

Phase 3: Days 46–90 — Automation Loop

• Build automated follow-up sequences for unbooked estimates: AI-driven SMS at 4 hours, email at 24 hours, AI voice callback at 72 hours
• Launch demographic-targeted marketing: direct mail to high-home-value zip codes within 6 miles of the warehouse, with messaging tuned to the household profile of those areas
• Enable AI-driven upsell prompts on the agent for known service plan customers and high-LTV demographic profiles
• Integrate full call data into Tideline’s CRM for ongoing analysis

Total implementation cost was modest. The real cost was Mark’s willingness to let an AI answer his customers’ phones — a non-trivial leap for a 23-year-old family business.

What Happened, Week by Week

Weeks 1–2: Quiet Wins After Hours

The first measurable change was after-hours. The AI voice agent was now answering 100% of calls between 6 p.m. and 7 a.m., plus weekends. Of those 916 quarterly after-hours calls (now annualized roughly to 305 in the first month), the AI booked 41% directly, scheduled callbacks for 27%, and gracefully referred or ended 32% (wrong number, out of area, non-plumbing inquiry).

In the first two weeks alone, the after-hours booking rate jumped from 12% (with messages) to 38% (real booked appointments). Mark called us on day 11. “I just got a 9:47 p.m. text confirmation for a Saturday-morning water-heater install I would have never seen until Monday. Do that thing again.”

Weeks 3–4: The Mid-Day Plug

When the AI moved into business hours and started handling overflow during the 9–11 a.m. window, the answer rate jumped from 67% in that window to 99%. Janelle described it as “finally being able to breathe.” Within three weeks, business-hours booking rate climbed from 41% to 49%, driven mostly by:

• Faster initial pickup (caller didn’t have time to hang up and call a competitor)
• Better triage (high-value callers got routed to the senior advisor directly)
• Pre-loaded caller context (the human advisor came onto the call already knowing it was a 2,100 sq ft home in a $480K neighborhood with a recurring service plan)

Weeks 5–6: Demographic Routing Pays Off

By week five, the demographic intelligence layer was producing observable revenue effects. Two examples that month:

• A call came in from a homeowner whose property data showed a 1972-built home in a flood-prone neighborhood and a recently expired service plan. The AI flagged the call as a re-engagement opportunity, routed to the senior advisor, and Tideline closed a $3,400 sump pump replacement + replumb of the basement drain line. The previous Tideline would have logged this as a routine pump call and quoted the $980 swap-out.

• A second call from a $1.1M home in the company’s prime zone, owned by a busy professional with two prior service tickets, was routed to the senior advisor and booked same-day for a tankless water heater install at $6,200. The advisor opened the call with: “Hi Mr. Chen, I see you’ve been with us since 2021 — we’ve taken care of the kitchen disposal and the upstairs toilet rebuild. What’s going on today?” The customer later told Janelle the call “felt like talking to a doctor’s office that actually remembered him.”

Personalization is not magic. It’s information used well. The system did the heavy lifting; the advisor delivered the moment.

Weeks 7–9: Follow-Up Compounds

The automated follow-up engine, by week seven, had started reviving leads that the old Tideline would have abandoned. Estimates given but not booked were now getting:

• An SMS recap with pricing and scheduling links within 4 hours
• An email at 24 hours with options to reschedule or modify
• An AI voice callback at 72 hours — yes, the AI calling the customer back, by name, with full context — to politely re-confirm or close

The unbooked-estimate close rate moved from 14% to 36% over a six-week window. That alone added roughly $74,000 in incremental revenue to the quarter — from leads Tideline had already paid to generate but was failing to convert.

Weeks 10–12: The Compounding Effect

By the end of week 12, the dashboard told a clear story:

• Inbound call volume: up 8% (some of this was natural seasonality; some was from the targeted direct mail to high-value zips that started in week 7)
• Answer rate: 99.4% overall (down from “we sometimes miss calls” to “we essentially never miss a call”)
• Booking rate on answered calls: 56% (up from 41%)
• Average ticket: $681 (up from $612), driven primarily by better routing of high-value calls to the senior advisor and AI-flagged upsell prompts on service-plan customers
• Total bookings: 38% higher than the previous 90-day window
• Estimated revenue captured: $2.37M vs. $1.71M baseline

The 90-Day Scorecard

Mark’s reaction in the week 13 review meeting was characteristically blunt.

“I wish I’d done this two years ago.”

What Actually Drove the Number

If you take nothing else from this case study, take this: the 38% wasn’t one big thing. It was four medium things stacking.

1. Stop missing calls. The simplest gain. Answering 99% of calls instead of 61% adds bookings even if nothing else changes.
2. Route the right calls to the right human. The senior advisor doesn’t waste time on $98 drain calls. The drain calls still get handled — just by the AI booking directly into the schedule. The advisor’s hours are spent on the high-value conversations he’s actually good at.
3. Use the data you didn’t know you had. Caller intelligence and demographic profiles aren’t a science project. They’re a context engine. Open every call with more information than the customer expects you to have, and conversion rises.
4. Stop letting estimates die. Automated follow-up at 4 hours, 24 hours, and 72 hours quietly resurrects a third of the revenue you used to throw away.

None of these are revolutionary individually. Compounding them in a single platform is.

How Caller Technologies Fits

Tideline’s stack now runs on a single platform — Caller Technologies — replacing what used to be a VoIP provider, an answering service, a CRM phone integration, and a spreadsheet of follow-ups Janelle kept on her desktop.

What’s running, simultaneously, under one system:

• AI voice agent answering every inbound call, 24/7
• Caller intelligence + demographic data — 2+ trillion data points, up to 150 attributes per caller, surfaced before the call connects
• Smart routing — emergency, high-value, plan customer, out-of-area, after-hours, and language-based rules
• VoIP phone system for the office
• Call analytics dashboards refreshed in real time
• AI coaching summaries on every human-handled call
• Automated follow-up sequences across SMS, email, and AI voice callback
• Marketing automation firing direct mail and digital campaigns based on call outcomes and demographic clusters

The integration is the product. Stitching together five vendors to do the same thing is what Tideline used to try. It didn’t work — not because any one tool was bad, but because the data never lived in one place at the same time.

Objection Handling

“Our customers want to talk to a real person.”

Tideline’s data tells the opposite story. Once the AI voice agent was indistinguishable from a human advisor in the first 30 seconds — and could actually book on the first call instead of taking a message — customer satisfaction scores went up, not down. The 82-year-old with a slow drain didn’t care if the voice was AI. She cared that someone answered, was patient, and got a tech to her by lunchtime.

“We’re not big enough for this.”

Tideline isn’t big. It’s a 23-person operation. Smaller shops gain proportionally more from AI voice because the cost of a missed call is a higher percentage of their revenue.

“The data feels invasive.”

Same publicly available property and demographic data used by direct mail, credit, and insurance industries for decades. No call recordings sold. No private medical or financial data exposed. The data simply makes the human conversation smarter.

“What if our techs hate it?”

Tideline’s techs love it. They show up to jobs with full context, the right tools, and an accurate scope. The AI weeded out the appointments that were going to be no-shows or out-of-scope before the truck rolled. Wasted truck rolls dropped by an estimated 19% in the 90-day window.

Conclusion

This wasn’t a moonshot transformation. Tideline didn’t pivot. It didn’t rebrand. It didn’t hire a Chief Data Officer. It simply stopped accepting that its phones were a problem it couldn’t solve.

Ninety days. Thirty-eight percent more bookings. Roughly two-thirds of a million dollars in additional captured revenue in a single quarter, with zero added trucks and zero added technicians.

If you’re running a plumbing company between $4M and $15M in revenue, the call-handling gap Tideline closed is almost certainly the same one sitting in your business today. The math will be slightly different. The structure will be the same.

Want to know what 90 days could look like for your plumbing company? Compare Caller Technologies against your current phone system — we’ll audit 30 days of your call data and show you, line by line, what’s slipping through, what’s worth fixing first, and what 38% looks like for your business.

Related reading

• AI Voice Agent vs Receptionist: Plumbing ROI Breakdown
• Treating Every Caller the Same Costs You Six Figures
• How HVAC Companies Lose Revenue After Hours (Fix It)

See the numbers for your own business with the ROI calculator, or compare plans on pricing.

See who’s calling before you say hello. The Caller Technologies AI voice agent answers 24/7, qualifies every caller with 150+ demographic signals — owner or renter, home value, income — and books real jobs while your crew works. Start your free trial — free until you book a paying job, no credit card.