An AI growth engine for home services.
A 3-truck HVAC company was bleeding revenue three ways — missed calls, slow callbacks, a dead customer database. One AI engine fixed all three. Cost per booked job fell from $310 to $156, and missed-call text-back alone recovered ~$40K a month in jobs that went to voicemail.
Figures blended across HVAC and plumbing lines. The model generalizes to roofing, electrical, garage doors, pest control, and solar.
A bucket with three holes.
Family-owned HVAC and plumbing: one metro, three trucks, ~$1.3M a year. The owner was the marketing department — paying $61 for shared leads four competitors also called.
~38% of calls hit voicemail. Leads sat 3.5 hours before a callback. ~6,000 past customers never heard from them again. We didn't sell ads — we built the operating system for the whole revenue loop.
- ~38% of calls to voicemail
- 3.5-hour average callbacks
- 6,000 past customers, never contacted
- 47 reviews at 4.1★
Seven subsystems, one loop.
Trigger → Capture → Convert → Maximize → Prove. Completed-job revenue flows back to the top, so the engine buys demand it can actually service.
Demand Radar
Weather, seasonality, and local signals pre-position spend ahead of the heat wave. Wired to the dispatch board: it throttles when trucks are full and bids harder for high-margin work.
Omnipresence Layer
Local Services Ads, Search, Google Business Profile, and Meta run as one front. Retargeting follows every caller who doesn't book; budget flows to profitable jobs, not cheap clicks.
Never-Miss Capture
Unanswered calls get a text-back in seconds; a 24/7 AI receptionist books the dispatch board at 2am. Answered-and-recovered rate: ~62% to ~98%.
Reputation Engine
A one-tap review link the moment the job completes. Unhappy signals route to a manager first; AI replies to 100% of reviews. More reviews → higher rank → more free calls.
Dispatch AI
Every lead qualified and booked in real time — 22-second speed-to-lead — then clustered geographically to fit more jobs per truck. Automated confirmations crush no-shows.
Revenue Maximizer
Post-job sequences turn one-time customers into members — $0 to $38K/month. Working ~6,000 dormant customers added ~$61K/month; financing and upsell context lifted average ticket from $480 to $710.
Command Center
One dashboard, closed-loop attribution — spend → booked → completed → revenue. Revenue-per-truck-hour, blended cost-per-lead, and membership ARR, live. Multi-location ready.
Seal the leaks first.
Before turning demand up, we stopped the bleed. Missed-call text-back alone recovered ~$40K a month in jobs the company was already paying to generate.
- Missed-call text-back in seconds
- 24/7 AI receptionist — books at 2am
- Photo-to-estimate intake
- Form, call, chat — one pipeline
- Answered & recovered: ~62% → ~98%
Same customer, five times the value.
Acquisition is expensive — the profit is in each customer. Sequences convert one-time jobs into members, and the AI works the ~6,000 dormant customers already on file.
- Memberships: $0 → $38K/month recurring
- Reactivation: ~$61K/month near-free revenue
- Financing prequalification
- Average ticket: $480 → $710
- Recurring floor for slow months
A flywheel in every job.
Competitors bolt six tools together. This is one system: each completed job becomes reviews, members, and reactivation revenue — making the next dollar of spend cheaper.
- Capacity-aware spend
- Review flywheel cheapens every channel
- Every call answered in seconds
- One system, not six tools
Why the numbers moved.
Seal the leaks, turn on demand, squeeze lifetime value — in that order. The closed loop is what compounds.
Why cost per booked job halved
Cheaper, higher-intent demand. Free organic calls from the review flywheel. Recovered calls that used to hit voicemail — more booked jobs from the same spend.
Why lead-to-booked nearly doubled
Speed and persistence: 22-second speed-to-lead, instant text-back, a 24/7 AI booker. Homeowners book whoever answers first — now that's always this company.
Why each customer became worth more
Memberships added a recurring floor. The reactivated database produced ~$61K/month of near-free revenue. Financing and upsell context lifted average ticket from $480 to $710.
“We thought we needed more leads. We were losing a third of our calls and had six thousand old customers we'd never called. They sealed the leaks — forty grand a month I didn't know I was losing — then turned the dial up. Now there's revenue even in the slow months. They built us a machine.”
Owner · residential HVAC & plumbing
Losing jobs you're already paying for?
Calls to voicemail. Leads sitting for hours. Past customers gone quiet. This system seals the leaks. Let's scope it.