Magic Voice family · Magic Receptionist
AI receptionist for HVAC companies. Picks up on the first ring, takes the info, captures the booking, sends an SMS confirmation, and emails the owner — 24/7.
White-label ready. Multi-tenant. Multi-vertical. Built on the Magic Voice stack.
Live · AI Receptionist Demo
Loading…
Audio runs in your browser. The agent answers as a generic AI receptionist — try booking an appointment or asking for a callback.
/hvac
We know what HVAC callers ask: no-heat in winter, no-cool in summer, gas smell, weird noise, tune-up scheduling, maintenance plan benefits. We capture equipment age, fuel type, and the actual symptom, route true emergencies to your on-call tech, and capture requested service slots for your dispatcher to confirm.
Five ways service businesses use Magic Receptionist
Pick the deployment that fits your business. We provision the forwarding rules, the AI script, and the after-call email — usually within a single business day.
Rollover when busy
$450 average estimate per recovered call. ~3 missed calls/day = $40K/year leaking.
Your existing line keeps its main role. When all desk staff are on calls, the carrier rolls overflow to the AI receptionist instead of voicemail. Caller gets a real intake; you get a booked job by morning.
Read the implementation guide →After-hours only
33% of body-shop call volume hits after 5pm. Most goes to voicemail. Most never calls back.
Set the AI to answer only after hours. Mornings start with a stack of qualified intakes — vehicle, damage, drivability, carrier — instead of a voicemail box you have to triage.
Read the implementation guide →Weekend coverage
Weekend collisions are the #1 trigger for new estimates Monday morning. Beat the competition by responding Saturday.
Weekend-only forwarding. Customers who get hit Saturday morning don't have to wait until Monday for the first response. The AI books the estimate, requests the photos, and pings the on-call estimator.
Read the implementation guide →Lead-gen ad campaigns
Standard click-to-call pages convert ~2-4%. With AI on the line, qualified-call-to-booking conversion jumps to 30-60%.
Every paid campaign gets its own forwarding number. The AI knows the campaign source, the offer, and the qualification questions. Every call ends with a structured email to your team — caller name, phone, qualification answers, and the campaign source attribution — so you can paste into your CRM. Direct CRM webhook export is on the public roadmap.
Read the implementation guide →Before-hours + after-hours
Customers calling before/after hours get a real conversation instead of a closed greeting. Same shift count; no expanded coverage.
The AI fills the gap on either side of your normal hours. Your existing staff handle the busy part of the day; the AI keeps the phones warm before opening and after close.
Read the implementation guide →Not sure which fits? — it'll ask about your call volume, your current handling, and recommend the right pilot.
How it works
Auto body, dental, legal, medical — each vertical ships with a script tuned to your callers.
Forward your existing number — or get a fresh one. We answer every ring, business hours or not.
Booking requests land in your team email with the caller's details, SMS confirmations go out to the caller, and the owner gets pinged on every call.
The math
Your phone rings. Nobody picks up. The caller hangs up, dials the next shop on Google, and books with them. Drag the dials and watch the number you're losing every year.
Math assumes 20% of missed callers would have booked. Adjust to your shop and the number still stings.
Annual revenue you're leaving on the table
$0
That's $0/month. Or about $0/week.
Stop the leak — see pricingNumbers are estimates based on the inputs above. Your actual leak depends on close rate + ticket mix + how fast competitors pick up.
Live wire
Anonymized snippets from across the network. Names changed. Outcomes real.
Sample transcripts. Names + locations are illustrative. Real network transcripts require tenant opt-in and full PII redaction.
Before · After
What happens after-hours when nobody's at the desk — with and without Magic Receptionist.
Tuesday evening at a service business (fictional).
Result: $850 booking lost. Customer gone for good.
Tuesday evening at a service business (fictional).
Result: $850 booking booked. Owner notified. Customer wow'd.
Pricing
Predictable monthly pricing. Each plan is one location, one AI agent — add another agent for ads, after-hours, or social channels. Fair use applies.
Agency or multi-shop operator? See the partner program — commission referral or wholesale per location.
Single-location AI receptionist. Answers, books, sends SMS confirmations.
500 min included · $0.45/min overage
300–1,100 min/mo (~10–35 calls/day)
Single location with after-hours coverage + busy-hour overflow.
1100 min included · $0.45/min overage
1,100–2,500 min/mo (~35–80 calls/day)
Annual
Annual prepay discount available
Ask about annual prepay terms during onboarding.
Fair use: Plans assume normal small-business call patterns. Sustained traffic above 3× the included minutes triggers a courtesy review and a tier upgrade conversation, not an immediate cutoff. We never throttle without telling you first. Read the full Acceptable Use Policy.
What HVAC owners ask
Pricing fits on a card; the rest doesn't. Real questions, real answers — the kind you'd ask another shop owner who already runs this.
An answering service takes a message — a person reads from a script, writes down the address and the words "no heat", and emails the dispatcher. You still call back, still re-ask the equipment questions, still try to figure out whether it was actually an emergency or could wait until tomorrow. Magic Receptionist actually finishes the work. The AI runs your tuned intake script the first time the caller picks up: equipment type, age, fuel source, exact symptom, urgency level. It captures routine service requests, applies maintenance-plan member benefits when the caller is on file, and pages your on-call tech for immediate callback on true emergencies (no-heat in winter, gas smell, ceiling drip). By the time the dispatcher sees it, it's already on the board.
Nothing breaks. Magic Receptionist answers every line in parallel — no hold queue, no busy signal, no voicemail box overflowing by 7:30 AM. Each caller gets first-ring pickup and the same intake flow. True no-heat emergencies page your on-call tech for immediate callback. Routine "my furnace is loud" or "can you do a tune-up next week" calls are captured with the requested service window and tech skill match noted. Maintenance plan members get priority slots automatically. Your office gets a structured queue of 30 bookings instead of a four-line meltdown — and the techs hit the road on real emergencies first.
We pre-load your service capabilities during onboarding — which equipment types you service, which fuel types (natural gas, propane, oil, electric, heat pump, mini-split), which residential vs light-commercial scopes, and any brand-specific certifications your team holds. When a caller mentions their equipment manufacturer, the AI confirms whether you service that type and gives an honest answer. The intake captures equipment age and visible nameplate info so the dispatched tech rolls with the right parts in the truck. Brand-specific deep-knowledge questions (warranty status, specific board recalls, refrigerant transitions) page a human or queue as a callback to your senior tech rather than getting bluffed.
Within the boundaries you set during onboarding. We load your standard diagnostic fee, after-hours fee structure, and any flat-rate service categories you publish openly. Callers asking "how much for a tune-up" or "what's your service-call fee" get the published rate. Anything that requires actual diagnosis ("how much to replace my compressor", "is this worth fixing or do I need a new system") gets a clear response: "Pricing on a repair like that depends on the actual diagnosis — let me get a tech out to take a look." The AI never invents pricing for work it can't accurately quote. That's the failure mode that turns into a complaint after the technician arrives.
Both stay. You don't change carriers, don't change main numbers, don't migrate your dispatch software, don't touch the trucks. Magic Receptionist sits as an overflow + after-hours + weekend layer: a simple call-forward rule routes unanswered and after-hours calls to your dedicated Magic Receptionist line. Calls your dispatcher answers during the day stay with your dispatcher. Booking requests land in your team email; live dispatch-software sync is on the public roadmap. Maintenance-plan members are loaded via CSV import during onboarding. We can switch the routing model during onboarding without touching the office.
FAQ
The 8 questions we hear most. Pricing, setup, emergency triage, security — all answered up front.
Magic Receptionist is an AI receptionist for HVAC companies. It answers calls 24/7, captures equipment age, fuel type, problem description, and urgency, captures service requests, flags after-hours emergency calls to your designated on-call tech, and pings the office on every booking. You keep your existing number — we sit on overflow and after-hours.
Plans start at $297/mo (Starter — 1 location, 500 minutes, $0.45/min overage). Growth is $597/mo (1 location with after-hours coverage, 1,100 minutes, $0.45/min overage). Add an extra agent on the same location for $49/mo. Reseller is custom for franchise groups and multi-location service operators. There is no free tier with metered cost — only paid plans plus per-minute overage and a fair-use policy.
We're in early access — setup is human-mediated. After you submit the form, our team confirms your setup, provisions your tenant, tunes the script with your hours, service area, fuel-type coverage, and emergency-fee structure, port-forwards your existing number to a Magic Receptionist line, runs two test calls with your dispatcher, and switches on overflow. We don't promise a specific go-live SLA in MVP — we work with you on timing. No new phone system, no rip-and-replace.
By default the AI takes overflow, after-hours, and weekend calls. New service requests get the full intake script (equipment age, problem, urgency, fuel type). Calls flagged emergency (no-heat in winter, no-cool in summer, gas-smell, water leak from indoor unit) flag for immediate callback and your on-call tech is paged. Maintenance plan members get priority routing per the rules set during onboarding.
Emergency triggers are tuned during onboarding — no-heat with outdoor temp below your threshold, no-cool with outdoor temp above your threshold, gas-smell language, water dripping from ceiling, frozen pipes risk. Those calls page your on-call tech for immediate callback. Routine maintenance, filter changes, and tune-up scheduling get next-available slots on your dispatch calendar without disturbing the on-call rotation.
Yes — during onboarding you provide your maintenance plan tiers, member benefits, and contract terms. When a caller mentions plan membership, the AI verifies against your member list (uploaded during onboarding or synced from your dispatch software), confirms benefit coverage, and applies the right pricing tier. Non-members hear your standard rate sheet plus the value pitch for joining the plan.
No long-term contract. Plans are month-to-month. Cancel anytime — your call data stays exportable for 30 days after cancellation. Annual plans get 2 months free if you prefer to lock in pricing. There are no setup fees on Starter or Pro. Service-area, rate-sheet, and emergency-trigger updates during onboarding are unlimited.
Yes. Call audio is processed through encrypted streams and stored encrypted at rest in Cloudflare R2. Transcripts are tenant-isolated in our D1 database with row-level access controls. Customer addresses, phone numbers, and payment hints are redacted from analytics aggregations. Sign-in to /admin uses WorkOS AuthKit.