Real math, not vibes. The actual dollar cost of a missed call, three-system comparison (voicemail, live virtual receptionist, AI receptionist), worked ROI examples for HVAC, plumbing, and roofing, when automation does NOT pay back, and an 8-question checklist for evaluating any vendor.
14-day free trial. No charge today. Guided setup available. Cancel anytime.
Quick answer
For a 1-to-10-truck home-service contractor missing 20 or more calls per month, yes. An AI receptionist at $49 to $349 per month typically recovers $7,000 to $12,000 in monthly expected revenue, for an annual net ROI of roughly $96,000 (HVAC) to $149,000 (roofing). The math does NOT pay back for long-cycle low-volume trades, owner-operators who personally take every call as differentiation, or shops with an unvalidated offer. Use the worked examples and 8-question vendor checklist below before committing.
The dollar cost
Short answer: a missed inbound service call costs between $250 and $1,200 in lost expected revenue, depending on trade and ticket size. For HVAC and plumbing emergencies, the high end is realistic because the caller dials the next contractor on Google within 60 seconds and books with whoever picks up.
Here is the calculation, built from publicly verifiable inputs.
Per U.S. Bureau of Labor Statistics Occupational Employment data and 2024-2025 industry benchmarks from Service Nation and ACCA (Air Conditioning Contractors of America), the average residential HVAC service ticket is $400 to $600 for diagnostic-plus-minor-repair and $5,000 to $12,000 for a system replacement. Plumbing service calls land between $300 and $700 (Angi 2024 cost data). Roofing repair tickets run $400 to $1,500, with full replacement jobs at $8,000 to $15,000 (HomeAdvisor and IBISWorld).
Inbound calls from homeowners actively Googling a contractor convert at 30 to 50%, far higher than form fills or web chats. Invoca's 2023 Contact Center Benchmark Report puts the average call-to-appointment conversion for home services at 39%.
A 2014 BIA/Kelsey study (still the most-cited research on this question) found that 85% of customers whose call goes unanswered will NOT call back. That number is the load-bearing stat for this industry, and it has held up in subsequent FCC and consumer-behavior studies because consumer patience for hold times has only gotten worse since smartphones made tapping the next result trivial.
For a 1-truck HVAC operator with a blended average ticket of $850 (mix of diagnostics, repairs, and occasional install lead):
For plumbing emergencies (burst pipe, no hot water, sewer backup), it is worse. Emergency tickets close at 55 to 70% because the homeowner is desperate, and the average emergency ticket runs $600 to $1,200. A missed emergency plumbing call costs roughly $400 to $700 in expected revenue, and roughly 95% of those callers do not call back. They book the next contractor on the SERP.
For roofing, missed-call cost is the highest per call because $8K to $15K replacement jobs originate from a $400 repair call. Volume is lower, but a missed roofing lead is a $600 to $1,500 expected-value miss including the replacement-conversion tail.
A 1-truck operator running modest paid-search and local SEO typically takes 80 to 200 calls per month. If 22% are missed (BIA/Kelsey SMB baseline, validated by Salesforce's 2024 State of Service report), that is 18 to 44 missed calls per month, or roughly $5,000 to $12,000 in monthly lost expected revenue. That is the number to put against the cost of any follow-up system you evaluate.
Why calls get missed
Before deciding whether automation pays back, you need to know which type of miss you are solving for. Not all missed calls are equal, and not all systems fix all five.
Reason 1
The biggest source of missed calls for owner-operators and small crews. Phone is in the truck, tech is 30 feet up a ladder or in a crawlspace.
Reason 2
Roughly 35 to 50% of plumbing and HVAC emergency calls fall outside 9-to-5 weekday hours. Most small contractors route after-hours to voicemail.
Reason 3
A Local Service Ads campaign, a Nextdoor mention, or a weather event can fire 5 to 10 calls into the same 30-minute window. A single-line front desk catches the first and drops the next four.
Reason 4
Single-receptionist shops have 30 to 90 minutes of daily coverage gaps. The phone routes to voicemail and stays there until someone notices.
Reason 5
Number-porting issues, GMB call-tracking misconfigurations, or a dead Twilio webhook. Rare but devastating because you do not know it is happening.
The first three (tech on a roof, after-hours, concurrent calls) account for 70 to 80% of missed calls in a typical 1-to-5-truck home-service business. They are also the three categories automation handles best, which is why the ROI math is aggressive in the small-contractor segment.
Three systems compared
There are realistically only three approaches a contractor can take to inbound calls they cannot personally answer. Each one has a different cost, a different recovery rate, and a different feel for the homeowner on the other end.
Free. Most homeowners hang up at the beep. A 1-truck shop missing 30 calls leaves $7,500 on the table every month.
Outsourced human, $1.00 to $1.50 per minute with $300 to $1,500 monthly minimums. Decent recovery, but rarely distinguishes a no-heat 95-degree-day emergency from an estimate request.
Trade-trained AI voice agent answers in under 2 rings, runs urgency triage, books direct to calendar. Starter is $49 for 100 calls. Pro is $149 for 400. Multi-Truck is $349 for 1,000. Overage $0.99 per call.
For a contractor doing fewer than 50 missed calls per month, AI is the clear ROI winner because the cost stays flat while voice quality and recovery rate stay high. Above 50 missed calls per month, the conversation shifts: a hybrid (AI handles overflow and after-hours, human handles complex or VIP) often wins on conversion AND customer perception. See the long-form AI vs traditional answering service comparison for the deeper feature-by-feature breakdown, or the AI answering service category page for the OnCrew take.
Worked examples
Same baseline assumptions for all three trades: 1 truck, modest paid + local SEO presence, 150 inbound calls per month, 22% missed-call rate (BIA/Kelsey baseline for SMB).
With AI receptionist (Pro tier, $149/mo)
With live virtual receptionist ($750/mo avg)
Both systems pay back. The AI system pays back roughly $27K per year more for this profile, mostly because the cost base is 5x lower and the recovery rate is higher. See the trade page for more on HVAC AI answering.
Same call volume (150/mo), but with a 50% emergency mix and a higher emergency ticket.
With AI receptionist (Pro tier, $149/mo)
Plumbing has the strongest ROI for automation of any trade because emergency callers will not, under any circumstances, leave a voicemail. Every emergency miss is essentially 100% lost without instant pickup. See the trade page for plumbing AI answering.
Roofing is trickier because the inbound call is often a repair lead that eventually becomes a $10K+ replacement job. Initial close rate is lower (these are estimates), but lifetime ticket value is much higher.
With AI receptionist (Pro tier, $149/mo)
Roofing ROI is dominated by the replacement-conversion tail. Even with a more conservative 6 to 8% repair-to-replacement assumption, the math still pays back at any reasonable assumption. See the trade page for roofing AI answering. Electrical contractors should review electrical AI answering for the panel-emergency triage equivalent.
| Trade | Net monthly ROI (AI receptionist) | Annual net ROI | Trade page |
|---|---|---|---|
| HVAC | $8,000 | $96,000 | HVAC AI answering |
| Plumbing | $8,621 | $103,452 | Plumbing AI answering |
| Roofing | $12,426 | $149,112 | Roofing AI answering |
These numbers assume conservative recovery rates and standard industry close rates. Your actual mileage will vary based on your average ticket, your close rate, and the quality of the system you choose. But the order of magnitude is correct: for a contractor missing 20+ calls per month, the ROI of automating follow-up is in the tens of thousands of dollars per year, against a cost in the hundreds.
What good looks like
"Automated follow-up" is a category that contains everything from a basic missed-call text-back autoresponder to a full conversational AI dispatcher. The cheap end of that range recovers maybe 25% of missed calls. The good end recovers 85 to 95%. The difference is in four specific capabilities.
The number-one variable in missed-call recovery is speed. A Lead Response Management study (MIT and InsideSales.com) found that contacting a web lead within 1 minute increases the odds of qualification by roughly 391% versus contacting within 5 minutes. The same physics apply to phone callbacks.
A message that just says "John called about his AC, his number is 555-1234" is barely useful. A real transcript captures the address, system age, symptom ("blowing warm air since 6am"), urgency ("baby in the house, 95 degrees"), and the preferred time window so a dispatcher can route the truck without a callback.
A good intake distinguishes between true emergency (no heat in January, no AC during a heat advisory, active water leak, no power, gas smell) and merely urgent (slow drain, intermittent issue, scheduled estimate). Emergencies get the next available slot or a live operator dispatch. Non-emergencies get the standard booking flow.
A real handoff means the appointment lands directly in the calendar your dispatcher is already watching (Google Calendar is the lowest-common-denominator nearly every small shop uses), with the urgency tag, address, customer phone, and a one-click link to the transcript. OnCrew specifically: Retell (voice engine) and Google Calendar are LIVE native integrations. ServiceTitan, HousecallPro, Jobber, and QuickBooks are currently assisted-setup (Zapier or webhook bridge configured on the onboarding call) with native API integrations targeting Q3 2026.
When it does NOT pay back
We have to be honest about this. There are real contractor profiles where automating missed-call follow-up does not generate enough ROI to justify even a $49 per month subscription.
Profile 1
A custom-cabinet shop or specialty remodeler taking 4 calls a week with a 90-day sales cycle does not benefit much from instant-pickup automation. The buyer is willing to wait, and voicemail-plus-callback works fine.
Profile 2
A handyman service at $75 per hour with 60% gross margin and 200 calls per month has thin economics. The math still slightly pays, but operational complexity eats into savings. A shared inbox plus a part-time scheduler often beats automation here.
Profile 3
Some owners build their brand on personally answering every call. If customers tell you "I called because YOU pick up," automation erodes your differentiation. Use it only for true overflow (after-hours, vacation, on-a-roof).
Profile 4
Facility managers, GCs, and utility dispatchers have established workflows requiring named contacts. AI is friction, not a feature. Stay human.
Profile 5
If your service mix, geography, or pricing is still in flux, automating intake locks you into a script before you know what should be on it. Run manual intake until your offer is dialed in, then automate.
If you fall into one of these five categories, the answer is stay manual, for now. Otherwise, the ROI math above is yours to capture.
Vendor checklist
If you have decided automation is worth piloting, here are the 8 questions that actually separate good providers from bad ones. Ask every vendor on your shortlist these questions, in writing.
A real number, not vague reassurance. A good AI provider will answer in under 2 rings (about 6 seconds). A good virtual receptionist will answer in under 3 rings. If they cannot tell you, they are not measuring it, which means it is bad.
Not a marketing demo. A real one. If they cannot, you will be the test market.
Specifically, what questions does your script ask to distinguish a "no heat in January" emergency from a "schedule an estimate" inquiry? If the answer is "we ask if it is an emergency," that is not triage, that is a checkbox.
Google Calendar? My CRM? An email? A vendor portal I have to log into? The right answer is directly in the calendar or CRM your dispatcher already uses, with one-click access to the transcript.
Insist on the truth: native API, Zapier bridge, manual handoff, or planned-for-future. Anything other than native API, live in production today is a workflow you will have to babysit. That is fine, as long as the vendor is honest about it.
Continuous coverage with the same script and same triage logic? Or does it route to a different queue with a different SLA? Most contractor revenue happens outside 9-to-5.
Cheap base plans with brutal overage are a common bait-and-switch. OnCrew, for example, is $0.99 per call overage with a real-time usage dashboard. Whatever your vendor is, get the number in writing and confirm the alert mechanism.
Can you leave on 30 days notice? Do you own your customer call recordings, transcripts, and call data, or does the vendor? If the answer to either is no, walk.
A vendor that answers these eight questions clearly and in writing is probably worth piloting. A vendor that dodges, hedges, or insists on a sales call before they will quote you is not.
Pricing snapshot
All ROI math above uses canonical OnCrew plan constants from this codebase, not stale prices. Overage is $0.99 per call on every tier, with a real-time usage dashboard.
Starter
$49/mo
100 included calls
$0.99 per call after
Solo contractors
Pro
$149/mo
400 included calls
$0.99 per call after
Growing contractor teams
Multi-Truck
$349/mo
1,000 included calls
$0.99 per call after
Multi-truck operations
See full plan details on the pricing page, or read the contractor answering service cost guide.
Quick answers for contractors deciding whether to automate inbound call handling.
The cheapest non-zero option is an AI receptionist on an entry-level plan. OnCrew's Starter tier is $49 per month for 100 calls, which covers most owner-operators doing 20 to 25 calls per week. A live virtual receptionist will run $300 to $900 per month minimum, so AI is roughly 6 to 10x cheaper at the entry level. The truly cheapest option is voicemail, which costs $0 and recovers about 12% of missed calls, so it is "cheap" only if you do not value the lost revenue.
Under 2 rings, ideally. The MIT and InsideSales lead-response research found that response times under 1 minute convert at roughly 391% the rate of responses 5 minutes later. For a live inbound call, that translates to picking up on the first or second ring, which is what a well-tuned AI agent does by default. Anything beyond 4 rings and you start losing the caller to the next Google result.
It depends on voice quality. As of 2026, the best AI voice agents (Retell, ElevenLabs Conversational AI, and similar engines) are convincing enough that most callers do not realize, especially over phone audio. Some providers (OnCrew included) let you disclose explicitly with a short opening like "this is the AI dispatcher for ACME HVAC, how can I help." Disclosure does not measurably hurt conversion in the data we have seen, and it builds trust.
Yes, often better than a live answering service, because there is no hold time and no agent looking up your emergency protocol. A well-configured AI intake asks the right triage questions (heat-related health risk, water leak severity, no-power duration), routes dispatch instantly, and texts the on-call tech. The risk to manage is rare edge cases (strong unhandled accent, poor cell connection, unusual call type). A good system has a "transfer to live" or "callback ASAP" fallback for those.
OnCrew specifically: today, the booking lands in Google Calendar natively, and we configure an assisted bridge (Zapier or webhook) into ServiceTitan, HousecallPro, Jobber, or QuickBooks during onboarding. Native API integrations for those four are on our roadmap for Q3 2026. Other providers vary widely. Ask every vendor for integration status in writing.
For a 1-truck HVAC operator taking 150 calls per month with a 22% missed-call rate, an AI receptionist at about $149 per month typically recovers $7,500 to $8,500 per month in expected revenue, for a net monthly ROI of roughly $8,000 and an annual net ROI of about $96,000. The biggest drivers of variance are average ticket size, close rate on inbound, and the percent of calls that are emergencies.
Missed-call text-back fires an SMS after the call has already been missed ("Sorry we missed your call, we will call you back shortly"). Recovery rate is maybe 25 to 40%. An AI receptionist prevents the miss by picking up live, running intake, and booking the appointment directly. Recovery rate is 80 to 95% because the call never gets missed.
No, and providers who pitch it that way are misleading you. AI handles inbound intake (the repetitive call-answering and scheduling part of the day). It does NOT replace dispatcher judgment, customer relationships, parts ordering, invoicing follow-up, or the other 30 things a good office manager does. Honest framing: it gives your office manager 10 to 20 hours per week back for higher-value work.
As of May 2026: Retell (voice engine) and Google Calendar are LIVE native integrations. ServiceTitan, HousecallPro, Jobber, and QuickBooks are currently assisted-setup, meaning we configure a Zapier or webhook bridge with you on the onboarding call. Native API integrations for those four platforms are planned for Q3 2026. We would rather state this plainly than overclaim.
For OnCrew specifically, a typical contractor goes from sign-up to live in 30 to 60 minutes if they already have their service area, pricing, and emergency criteria documented. If you need to write your intake script from scratch, plan for 2 to 3 hours total over a couple of days. Other vendors range from same-day to 2-to-3-week implementations depending on integration complexity.
Every reputable provider gives you the full transcript and recording, so you can hear exactly what happened. If the AI mishandled it (wrong slot, missed urgency cue, wrong pricing), you flag the call in the dashboard, the provider reviews it, and the script gets updated. If a vendor cannot give you recording, transcript, and a feedback loop, that is a red flag.
Pull a call report from your phone system (Google Voice, RingCentral, OpenPhone, Twilio, etc.) for the last 60 days. Count the calls that went to voicemail or had no answer. Divide by total inbound calls. That is your missed-call rate. If you cannot get a call report from your current phone system, that is itself a sign you are flying blind on the most important number in your business. OnCrew has a free missed-call calculator at /tools/missed-call-calculator that walks through this if you want a quick sanity check.
Keep evaluating
The two calculators plug your real numbers into the formulas above. The buyer's guide and the missed-call playbook turn the math into a setup pass.
Plug your own call volume, close rate, and average ticket into the worked-example formulas. Returns monthly and annualized recovery estimates.
Run the calculatorCompare voicemail, live receptionist, hybrid, and OnCrew side by side at your call volume. Uses the same per-minute and per-included-call assumptions cited above.
Compare costsEight evaluation criteria, six red flags, AI vs live vs voicemail matrix, trade fit, and a full vendor shortlist. Pair it with the checklist on this page.
Read the buyer's guideThe companion methodology page. Five-trade planning scenarios, public BLS wage and IRS FICA sources, and citation-ready snippets for partner blogs.
Read the benchmarkSix-step playbook for contractors tired of losing jobs to voicemail. Audit call flow, size up cost, pick coverage, set urgency rules, forward and test, review weekly.
Run the playbookThe OnCrew take on the AI receptionist category. Live in production, 24/7 pickup, urgency triage, and Google Calendar handoff for HVAC, plumbing, electrical, and roofing shops.
See the category pageBrowse every guide on the contractor resources hub.
The math on missed-call follow-up is not subtle. For most 1-to-10 truck home-service contractors, automating intake recovers an order of magnitude more revenue than it costs. Start a 14-day free trial on your existing number, or run the calculators with one month of your call-log data first.
14-day free trial. No charge today. Guided setup available. Cancel anytime.