AI Receptionist for HVAC Companies

Most HVAC calls split into three buckets: true emergencies (no heat in winter, no AC in extreme heat, gas smell, water damage), urgent-but-not-life-safety (system not cooling well, strange noises, intermittent failures), and routine maintenance. A human dispatcher mentally triages these in about 30 seconds. The Prisma Voices AI does it in under 5 — by listening for keywords in your configured emergency dictionary, asking 1 to 2 follow-up safety questions, and pulling up your on-call schedule.

When a caller says 'I smell gas,' the AI recognizes that as a hard-stop emergency, instructs the customer to evacuate and call the gas utility first, then pages your on-call tech with full caller details. When a caller says 'my AC is making a clicking noise,' the AI books a same-week diagnostic. The dispatching logic is your business — the AI just executes it consistently every time.

Every HVAC company knows the pattern: the first 90°F day in May triggers 5x the usual call volume, and the first hard freeze in November does the same. Most shops handle this by paying overtime, hiring temps, or accepting that 30 to 40% of those calls go to voicemail. Prisma Voices answers every one of them simultaneously — there is no ceiling on concurrent calls, no hold queue, no 'press 1 for service.'

The math is dramatic. A shop that books 20 jobs from 50 peak-day calls (40% capture) becomes a shop that books 35 to 40 jobs from those same calls (75 to 80% capture, with the rest being misdials and existing customer questions). At an average ticket of $400, that is $6,000 to $8,000 in additional revenue per peak day, captured automatically.

The AI does not replace your dispatch software — it feeds it. Every booked call lands on your Google Calendar (or Cal.com) with a structured event: customer name, phone, full address, service issue description, equipment age and type, recommended technician based on your zone rules, and a flag for any escalation triggers. Your dispatcher's morning routine becomes 'review the AI's bookings and assign trucks' instead of 'play voicemail roulette and chase callbacks.'

For shops on ServiceTitan, Housecall Pro, FieldEdge, or Jobber, the calendar sync is bidirectional: technician availability flows in (so the AI never books over a busy slot), and new bookings flow out into your existing job board. Your team keeps using the dispatch software they already know.

After 6 PM, on weekends, on holidays — when your office is dark — the AI's behavior changes based on your configuration. For routine issues, it offers next-business-day slots and books them. For configured emergencies, it pages your on-call tech via SMS within 30 seconds, including a one-line summary ('No heat call from 412 Oak St, 70-year-old caller, furnace dead, baby in the home'). Your on-call tech calls back from a real human number — the AI hands them a fully briefed customer instead of a voicemail to decode.

If your on-call rotation is complex (e.g., commercial accounts go to Tony, residential to Mike, refrigeration to Sarah), configure those rules once and the AI follows them every night without exception. No more 'I forgot it was my weekend' incidents.

Maintenance plans are recurring revenue, and the best shops convert 30 to 50% of repair calls into plan members. Most lose this revenue because dispatchers are too busy to pitch the plan or forget the script. The AI never forgets. Configure your plan terms once and the AI mentions them at the right moments: after booking a repair, when a tune-up is due, when the customer asks about pricing for a new install. It quotes the plan accurately, answers objections from a configurable script, and routes the contract for SMS signature on the spot.