Rank every no-heat call by outdoor temperature and who is actually in the house, using thresholds you set before the cold snap rather than during it.
An AI agent for no-heat call triage is a 24/7 digital assistant that answers winter furnace calls, records the indoor and outdoor temperature and who is living in the property, ranks the call against thresholds the owner configured, and either pages the on-call technician or books the earliest slot that fits. It stops a house at 44F with a newborn in it from sitting behind four thermostat questions in the same morning queue. In a cold snap an HVAC company can take a week of calls in a day, and every one of them says the heat is out. What separates them is risk, and risk is a function of two things the caller can answer in ten seconds: how cold it is outside, and who is inside. The agent asks both on every call. A property below your configured indoor floor with an infant, an elderly resident or someone with a medical condition is escalated. A well-insulated home at 58F with two adults is booked. The thresholds are yours; the agent does not invent them.
Collects the two facts that set priority, then routes on your thresholds.
Answers no-heat calls day or night without putting anyone in a hold queue
Asks the current indoor temperature and reads the outdoor temperature for the address
Asks who is in the house: infants, elderly residents, anyone with a medical condition
Applies your configured thresholds, for example under 55F indoors with an infant escalates
Gives only the approved holding advice — check the thermostat setting and the furnace switch
Pages the on-call technician for at-risk homes and books the rest into ranked order
During a freeze the constraint is never demand, it is crew hours, and the only useful question is who gets the first van. Answered by a phone queue, that question resolves to whoever redialled most persistently. Answered by this agent, it resolves to the household with a 46F living room and a ninety-year-old in it. The thresholds are written by you in advance, applied identically to every caller, and visible in the log afterwards — which also means you can defend the dispatch order if someone asks why they waited.
A simple, three-step flow.
The agent picks up on the first ring, confirms the service address and pulls the current outdoor temperature for that location.
It asks what the thermostat reads now, whether the system is blowing cold air or nothing, and who is in the house — infants, elderly residents, medical conditions.
The answers are tested against your thresholds. At-risk homes page the on-call technician; the rest are booked in ranked order with a realistic window, and everything is logged.
A realistic use case with concrete timing and output.
Scenario: a 7-technician HVAC company during a three-day freeze with outdoor lows of 12F. Normal winter volume is 30 calls a day; on the first cold morning it hits 96. The agent answers all of them. The owner's thresholds are set to escalate below 55F indoors with an infant under two, a resident over 75, or a documented medical condition, and below 45F indoors regardless of occupancy. Twelve calls clear those thresholds and go into the emergency queue in order of severity, the coldest first — a 41F house with an eighty-two-year-old resident is the first van of the day at 7:20am. Thirty-one callers turn out to have a thermostat on Cool or a tripped furnace switch and are resolved on the call. The remaining fifty-three are booked across three days with an honest wait quoted at the time, not promised same-day and then missed.
Roles that gain practical value from this AI agent.
A freeze is your highest-revenue and highest-risk week, and dispatch order decides how both go.
You get a ranked queue with temperatures and occupancy attached instead of ninety identical no-heat tickets.
The escalation thresholds become a written standard you can train to and audit afterwards.
Habitability rules turn on indoor temperature, and the agent captures it on the call in the tenant's own words.
A no-heat call that is a thermostat setting should never consume a van slot in the middle of a cold snap.
You arrive knowing the indoor temperature, the system behavior and who is in the house before you knock.
Key tools and what the AI agent does inside each.
Answers every no-heat call on your main line, including the surge volume during a freeze.
Pages the on-call technician for at-risk homes and texts booked callers their window and position.
Creates the no-heat job with indoor temperature and occupancy in the notes so the board reflects real priority.
Books non-escalated calls into genuine open slots across the crew rather than overbooking the day.
Logs every call with its temperature reading, occupancy answer and dispatch decision for post-freeze review.
Six practical scenarios that this AI agent excels in.
Common questions about using the AI agent in workflows.
An AI agent for no-heat call triage is a 24/7 digital assistant that answers winter furnace calls, records the indoor and outdoor temperature and who is living in the property, ranks the call against thresholds the owner configured, and either pages the on-call technician or books the earliest slot that fits. It stops a house at 44F with a newborn in it from sitting behind four thermostat questions in the same morning queue. Unlike a voicemail box or an answering service, it collects the facts that set priority and produces a ranked dispatch decision rather than a message.
Most companies escalate below 55F indoors when there is an infant under two, a resident over roughly 75, or anyone with a documented medical condition, and below 45F indoors regardless of who is home. Some add an automatic escalation when the outdoor temperature is below freezing and the system is producing no heat at all. The numbers are yours to set; the agent only applies them.
Because the same indoor temperature is a different risk in different households. A 50F house with two healthy adults is uncomfortable; the same house with a newborn or a ninety-year-old is a genuine medical concern. The occupancy answer changes the priority independently of the fault, which is why it is asked on every call and not only when the caller volunteers it.
No. The temperature and occupancy thresholds are a fixed configuration, and the agent applies them literally however insistent the caller is. It will not move someone up the queue, waive an after-hours fee or promise a van that dispatch has not authorized. If the caller disputes the ranking, the call goes to your on-call human, who is the only party able to override it.
Only what you approve, and it is short: check that the thermostat is set to heat and above room temperature, and check that the furnace service switch and breaker have not been knocked off. It does not talk anyone through relighting a pilot, opening a burner compartment or resetting a lockout, and it never suggests improvised heating that involves an open flame or a cooker.
Triage stops immediately. The agent switches to the gas safety script: leave the building now, do not touch light switches, call the gas utility's emergency line and 911 from outside. It does not continue asking about the furnace, and the on-call technician is paged at the same time. That branch overrides everything else on the call.
Honestly. When the diary is full the agent quotes the real next window rather than same-day, tells the caller where they sit relative to the escalation criteria, and offers to text them if an earlier slot opens. Setting an accurate expectation at 8am produces far fewer angry calls at 4pm than a same-day promise nobody can keep.
Rank every no-heat call by outdoor temperature and who is actually in the house, using thresholds you set before the cold snap rather than during it.