AI Receptionist: Plumbers, Dentists, HVAC Decision Tree

15 min read By FindSkill Team

85% of unanswered calls don't call back. The 60-minute weekend missed-call audit and build/buy/wait decision tree for plumbing, dental, and HVAC businesses in 2026.

Table of Contents

If you run a plumbing business, a dental practice, or an HVAC company, here’s the math that should keep you up Sunday night. Two recent industry datasets — NextPhone’s 2026 analysis of 347,000 contractor calls and CallJolt’s 2026 missed-call breakdown — converge on the same picture: 74% of calls to small contractors go unanswered, 62% of callers won’t leave a voicemail, and 85% won’t try again if they don’t reach a human. Zendesk’s broader CX research puts the no-callback number at 93% across industries, so the directional range is “about 4 in 5 callers never call back.”

The CallJolt dataset breaks the unanswered fraction out by trade, which is the part nobody quotes: HVAC 64%, plumbing 61%, electrical 58%, roofing 67%. Dental practices fare a little better — Goodcall’s 2025 dental data shows ~35% of incoming calls missed — but the dollar leak is comparable because each lost dental patient is a multi-year relationship, not a one-off invoice. For a plumber doing 42 calls a month, that’s roughly 26 calls hitting voicemail and 22 callers never trying you again. At an average ticket of $400 for plumbing repair or $2,000 for an HVAC service call, the missed-call leak is real money — usually somewhere between $5,000 and $20,000 a month, with Resonate AI putting the average loss per individual missed call at around $450.

This post is for owners of plumbing, dental, and HVAC businesses who’ve heard about AI receptionists and want to know whether to build one, buy one, or wait. There’s no single right answer; there’s a decision tree based on your Monday-morning voicemail count and three other inputs. We’ll walk all three trades through the same tree.

The 60-minute weekend missed-call audit

Run this Sunday evening or Monday morning. The goal is to know your real number before any vendor demo wastes your time.

Step 1 — Pull the last 30 days of call logs. Most VoIP and phone systems (RingCentral, Vonage, NextPhone, your traditional phone company) export this. You want: total calls received, calls answered live, calls that went to voicemail, and average ring duration before pickup.

Step 2 — Count the after-hours fraction. What percent of your calls came in outside 8am-5pm Monday-Friday? For plumbers, this is usually 25-40% (emergency calls don’t respect business hours). For HVAC, 30-45% (especially during heat waves and cold snaps). For dental practices, 15-25% (post-procedure pain calls and same-day cancellations from the day before).

Step 3 — Bucket your missed calls into 4 categories.

  • Category A: Emergency/urgent (“My water heater is leaking right now,” “My AC is out and it’s 95°F,” “I think I cracked a tooth”). High value. High urgency. The category where missed calls hurt most.
  • Category B: Booking inquiries (“I’d like to schedule a teeth cleaning,” “Can someone come look at my furnace next week”). Medium urgency. Most likely to call a competitor next.
  • Category C: Existing customer follow-ups (“Just confirming my appointment tomorrow,” “Question about my invoice”). Lower urgency but high impact on retention.
  • Category D: Sales / vendor calls / wrong numbers. Low value. The fact that they’re missed is usually fine.

Step 4 — Estimate the revenue per category. Take your average ticket size for each business line and multiply by your typical close rate when a call is answered live. Most service trades convert 30-60% of inbound calls into work orders.

Step 5 — Compute the leak. Missed calls × close rate × average ticket = monthly leak. For a typical 6-truck plumbing operation, the leak is usually $8,000-$25,000/month. For a 3-chair dental practice, $5,000-$15,000/month in retained-customer revenue. For a 2-truck HVAC contractor, $10,000-$30,000/month during peak season.

That number is your decision input. Without it, you’re flying blind on whether to spend $50/month or $300/month on a solution.

The build / buy / wait decision tree

One tree, three vertical-specific examples.

START → Did the audit show <$2,000/month leak?
        ├── YES → WAIT. Re-audit in 6 months.
        └── NO → Continue.
                 │
                 ├── Is leak between $2,000-$8,000/month?
                 │   ├── YES → BUY. Pick a $50-$150/month plan.
                 │   └── NO → Continue.
                 │
                 ├── Is leak between $8,000-$20,000/month?
                 │   ├── YES → BUY. Pick a $200-$400/month plan.
                 │   └── NO → Continue.
                 │
                 └── Is leak >$20,000/month with high call volume?
                     ├── YES → BUY HYBRID. AI + human escalation,
                     │         $400-$1,500/month.
                     └── BUILD only if you have:
                         - In-house developer time
                         - Need for deep CRM integration
                         - Compliance posture (HIPAA for dental)
                         - 6+ months of timeline

The default answer for almost every plumber, dental practice, and HVAC contractor is BUY. Building takes engineering time most service businesses don’t have. The vendor market is mature enough in 2026 that off-the-shelf options handle 80% of use cases at a fraction of build cost.

Vertical 1 — Plumbing businesses

The pattern: 30-40% of calls are after-hours emergencies. Most callers don’t have a relationship with your shop yet — they Googled “plumber near me” five minutes ago. The first business that picks up wins the job.

The vendors that actually fit:

  • Goodcall ($95-$199/mo) — purpose-built for service businesses. Drag-and-drop workflow builder. Handles emergency triage (“if caller says ’leaking right now,’ route to dispatch immediately”).
  • Rosie (~$199-$299/mo unlimited) — emotion recognition built in. Strong for home-services where caller stress is part of the qualification.
  • Voiceflow (build-your-own, ~$50/mo plus dev time) — only if you have someone who’ll maintain it.

The plumbing-specific question to test in any vendor demo: “How does the AI handle ‘my water heater is leaking right now’?” The answer should be: immediate dispatch escalation, customer-info captured in 60 seconds, optional SMS confirmation while the technician is en route. If the demo doesn’t handle that flow cleanly, the vendor isn’t ready for plumbing.

Build vs. buy for plumbers: buy. You don’t have time to maintain prompt logic between burst-pipe season and HVAC overlap season. The exception: if you’re a 15+ truck operation with a dedicated office manager doing dispatch and CRM integration is the bottleneck, the cost-per-call math may flip toward a custom build.

Vertical 2 — Dental practices

The pattern: 15-25% after-hours, but a different shape — most calls are existing-patient follow-ups and same-day cancellations rather than emergencies. The hidden leak is the empty chair on Tuesday afternoon when someone canceled Monday at 5:01pm and the office didn’t have time to refill the slot.

Vendors that ACTUALLY have HIPAA / BAA tiers for dental:

  • RingCentral AIR Pro for Healthcare (launched March 2026) — explicitly HIPAA-ready, sits under RingCentral’s existing BAA framework when configured properly. The cleanest fit if you’re already on RingCentral or willing to migrate.
  • Goodcall ($95-$199/mo) — has a healthcare tier with BAA available; confirm the recordings/transcripts scope before signing.

Vendors that are NOT safe for dental (PHI calls forbidden):

  • Smith.ai — Smith.ai’s own FAQ explicitly states they are not HIPAA-compliant and cannot handle calls involving PHI. Don’t use them for dental front-desk work.
  • Rosie — third-party comparison guides (Voksha 2026) confirm Rosie is not HIPAA-compliant. Same deal.
  • Aira / Allo / LeadTruffle / NextPhone — no public BAA documentation as of May 2026. Treat as non-HIPAA unless the vendor offers a private BAA on contract.

HIPAA is the constraint that changes the buy decision. Any AI receptionist for a dental practice must sign a Business Associate Agreement (BAA) and process patient information in a HIPAA-aligned way. Do not skip this step. Confirm the BAA before the demo, in writing. The shortlist that holds up in May 2026 is RingCentral AIR Pro for Healthcare and Goodcall (healthcare tier).

Goodcall’s dental-specific data shows roughly 35% of dental calls go unanswered — lower than the home-services trades but enough to drive 5-15K/month in gift-card and patient-retention leakage at typical practice sizes.

The dental-specific question to test: “How do you handle ‘I need to cancel my Tuesday cleaning’?” The answer should be: cancellation captured, slot returned to your scheduling system in real-time, optional waitlist callout to fill the gap. If the AI just takes a message and emails it to the front desk for Tuesday morning, you’ve lost the slot.

Build vs. buy for dental: buy. The HIPAA BAA framework is too much to handle in-house unless you’re a multi-practice dental service organization. For a single-practice or small-DSO buyer, the off-the-shelf vendors with existing BAAs win on time-to-deploy alone.

Vertical 3 — HVAC contractors

The pattern: the most seasonal of the three. Summer heat waves and winter cold snaps drive call volume up 5-10x baseline. Off-season, you may be at 20 calls/month; peak season, 200+. The AI receptionist needs to scale without your monthly fee scaling proportionally.

Vendors that handle HVAC seasonality:

  • Goodcall — flat-rate pricing ($95-$199/mo) handles seasonal spikes without per-call surcharges that bite during heat waves.
  • NextPhone ($199/mo unlimited) — same flat-rate logic; unlimited calls means peak-season volume doesn’t change your bill.
  • Trillet (starts $49/mo, then $0.20/minute) — cheaper at baseline but the per-minute math gets ugly during summer. Avoid for heavy seasonal volume.
  • Allo / WithAllo / LeadTruffle — purpose-built for HVAC and home-services contractors with lead-qualification scoring built in.

The HVAC-specific question to test: “How do you handle ‘My AC is out and the house is 90°F with my baby in it’?” Emergency-tier dispatch, customer information captured fast, technician routing based on geographic proximity to dispatch. Bonus: does the AI track emergency calls separately so you can prioritize them in your job-board queue? The good vendors do this.

Build vs. buy for HVAC: buy, with one caveat. If you operate across multiple states with regional crews and your dispatch logic is genuinely complex (which technician, which geographic zone, which equipment expertise), Voiceflow-built or even Air-by-RingCentral might be worth evaluating because the customization depth is higher. For single-state, single-region HVAC contractors, off-the-shelf wins.

What this means for you

If your audit shows under $2,000/month leak: Wait. Run the audit again in 6 months. The vendor market is competitive and prices are dropping; what costs $200/month today will likely cost $100/month in 12-18 months for equivalent features.

If your audit shows $2,000-$8,000/month leak: Buy a $50-$150/month plan. Goodcall, Aira, or Trillet (for plumbers/HVAC) — NextPhone or Aira (for dental). Sign a month-to-month if available; don’t lock into annual contracts before 90 days of real usage data.

If your audit shows $8,000-$20,000/month leak: Buy a $200-$400/month plan with full features. Test 2-3 vendors in parallel for 30 days. Most of them have free trials. The one that handles your top 5 call types cleanly wins.

If your audit shows $20,000+/month leak: Hybrid AI + human escalation. Smith.ai is the cleanest fit; Allo and LeadTruffle for HVAC and home-services contractors specifically. The $400-$1,500/month cost is justified at this leak level by even modest improvement.

If you’re already on a phone system you can’t easily change: RingCentral users can layer AIR (built-in AI receptionist) without migration. Vonage users have similar in-system options. Don’t migrate phone systems just to get an AI receptionist; layer it on top of what you have.

What it can’t do

It can’t handle truly complex dispatch decisions. A plumber’s AI can route an emergency call to the on-call tech, but it can’t decide which emergency takes priority when three come in within 10 minutes. That’s still dispatch’s job. Don’t let a vendor demo convince you AI replaces your dispatcher.

It can’t substitute for clinical judgment in dental. The AI handles booking, cancellations, insurance-eligibility-checks, and basic FAQ. It does not handle “should I come in or take ibuprofen and call you in the morning?” That’s an actual clinical question and needs a real person.

It can’t cover the relationship calls. Existing customers calling to check in, wanting to talk to “Jim who fixed our furnace last year,” or just wanting reassurance — those calls need a human. Make sure your AI has a clean handoff to a real number for these.

It can’t fix a bad CRM. If your job-tracking is a paper notebook, the AI can capture the call but can’t put it into your workflow. Get the CRM clean first, or pick a vendor with integrations to whatever you’re using (ServiceTitan, Housecall Pro, Jobber, Dentrix, Open Dental).

It can’t bypass licensing for medical advice. Dental practices: be careful with any AI feature that gets close to “what should I do about my tooth pain?” Even routine triage advice can cross into unlicensed-practice territory in some states. Stick to scheduling, insurance, and “we’ll have the dentist call you back.”

The bottom line

For most plumbing, dental, and HVAC businesses, the answer is buy, not build. Run the audit, find your real number, then pick a vendor in the $50-$300/month range that fits your call pattern. The 85% don’t-call-back number is the math that justifies the spend.

Three things to do this week:

  1. Pull the last 30 days of call logs and run the 60-minute audit
  2. Pick 2 vendors that match your vertical and request 30-day trials
  3. Have the trial AIs handle real calls — not demo calls — for at least two weeks before signing anything longer than month-to-month

If you want the full local-services AI playbook — including the dispatch-integration patterns, the CRM-handoff scripts, and the customer-experience design for trades businesses — that’s covered in the trades-electrical-plumbing course and the dentists course.

Companion read: this is a sister-vertical to the auto-shop AIR audit we covered last week — same decision tree, different industry.

v1.1 note (2026-05-10 evening): This post was updated within hours of initial publication after Perplexity Pro research surfaced two material corrections: (1) Smith.ai’s own FAQ explicitly states the service is not HIPAA-compliant — it was originally listed as a viable dental option, which was wrong; the dental shortlist now reads RingCentral AIR Pro for Healthcare + Goodcall (healthcare tier). (2) The “85% never call back” headline now sits inside the broader 78-93% range from NextPhone (347K-call dataset), CallJolt 2026, and Zendesk CX research, with trade-specific unanswered-call rates (HVAC 64%, plumbing 61%, electrical 58%, roofing 67%, dental 35%) added. If you read the earlier version, please re-read the dental section.

Sources

Build Real AI Skills

Step-by-step courses with quizzes and certificates for your resume