One phone line, two businesses with opposite clocks
A standby generator company answers two kinds of call on the same number.
The first is an installation enquiry. Somebody wants whole-home backup power. That is a project: a load assessment, a proposal, a gas line, an electrician, a transfer switch, permits, a utility or gas inspection, and equipment that may be weeks out. From first call to commissioning is measured in weeks, sometimes months. A same-day callback is excellent service on that call.
The second is a service call. A customer's unit did not start, or it started and shut down, or the controller is showing a fault, and frequently the power is out right now. That is an emergency measured in hours. A callback tomorrow is not service; it is a lost customer telling their neighbours.
As of September 2026, the single most consequential operational decision a generator company makes is whether those two calls are handled by the same process. Most are. The project cadence wins, because it is the majority of the volume in fair weather, and then the emergency callers get project-speed treatment at exactly the moment they are least able to tolerate it.
This guide is about separating them at the front door, then building the rest — contracts, monitoring, permits, the storm wall — around both. The operational system underneath is Run with Jarvis, and its natural-language control layer is described in the Jarvis AI Brain overview.
Triage at the front door
Triage is one question asked early: do you have power right now, and is this an existing unit?
Those two answers sort every caller into one of four buckets, and each bucket has a different correct next action.
Existing unit, no power. Emergency. Capture make, model, symptom, fuel type, whether it cranks, what the controller shows, and dispatch or escalate immediately. This caller does not get a callback tomorrow.
Existing unit, power is fine. Service or maintenance. A fault light, a failed exercise cycle, a due service. Book it into a normal slot with the model and history attached.
No unit, wants one. Installation lead. Capture the property type, panel location, fuel availability, what they want to back up and their timeline, give a tier range, book an assessment.
Not our unit, wants us to take it over. The most commercially interesting caller and the one most often mishandled — an orphaned generator whose original installer is gone or unresponsive. They are asking to become a maintenance customer. Capture the model, age and service history, and treat it as both a service call and a contract lead.
Asking the triage question first is what makes this work, and it is exactly the kind of consistency a human answering between site visits cannot maintain. The platform's AI receptionist runs the same triage on every call, in English or Spanish, at 2 AM as reliably as at 2 PM. Where that matters most is after hours, because a generator emergency is disproportionately an after-hours event — the operational shape of that is in the after-hours calls playbook.
The sizing conversation you can and cannot have
A homeowner asking "how much for a generator?" is really asking two things: is this a five-thousand-dollar decision or a twenty-thousand-dollar one, and what does it involve.
Both are answerable on the phone without over-promising.
The tier range is answerable because the market has natural bands: a small air-cooled unit backing up essential circuits, a larger air-cooled unit carrying most of a typical home, and a liquid-cooled unit for a large property or one with heavy loads. A caller who learns which band they are probably in has enough to decide whether to book an assessment.
What cannot be answered on the phone is the actual sizing, and the honest reason is worth saying out loud: the load calculation depends on the electrical panel, the specific circuits the customer wants to carry, the gas supply available at the meter, and the local code. Anyone quoting a firm size and price from a phone description is guessing, and the guess surfaces as either an undersized unit that trips under load or an oversized one the customer overpaid for.
What the intake should capture is the input list for that assessment: square footage, panel age and location, propane or natural gas, whether the customer wants essential circuits or whole-home, whether there is air conditioning to carry, well pump, electric range, electric vehicle charging. Those decide which assessment slot and which technician, and they let the proposal start from real information.
Fuel type deserves more weight at intake than it usually gets, because it changes the project rather than just a line on the quote. Natural gas depends on what the meter and the service line can actually deliver once the generator's demand is added to the house's existing appliances, which sometimes means a utility upgrade with its own timeline outside your control. Propane depends on tank size, placement, regulator capacity and refill logistics, and on a customer accepting a tank in their yard. A caller who says "gas" may mean either, and the two produce different proposals, different permits and different lead times. Asking which one, and whether a tank already exists, costs one question and prevents a proposal built on the wrong assumption.
The coordination complexity is the part customers do not anticipate: a generator install typically needs both a gas connection and electrical work, which means either two trades or one company with both licences, plus a permit and often a utility or gas inspection. Sequencing that is project management, and the customer should hear the sequence at the assessment rather than discovering it in week three. Where electrical scope is subcontracted or shared, the coordination pattern overlaps heavily with the electrical contractor operations guide.
Maintenance agreements are the whole business model
An installation is a one-time transaction. A maintenance agreement is an annuity, and in generator work it is also the thing that makes the product actually work.
A standby generator is a machine that sits unused for months and is then asked to start instantly under load. That is a demanding duty cycle, and it is why manufacturers specify annual service — oil, filters, plugs, coolant on liquid-cooled units, battery check, load test, and verification that the weekly exercise cycle is running. Skip it for three years and the unit that the customer paid five figures for does not start on the night it matters.
Commercially, agreements do four things at once. They produce predictable scheduled revenue. They put a technician in front of the unit annually, which surfaces the failing battery before the outage. They dramatically reduce emergency callouts, which are the least profitable and most disruptive work you do. And they create the customer record that makes every future call fast.
Operationally they need three things from the software. A cadence the calendar holds — annual, semi-annual, or per manufacturer spec — without a human maintaining a spreadsheet. A customer record that remembers the unit: model, serial, install date, fuel type, transfer switch, battery replacement date, last service. And a renewal that runs itself, because an agreement that quietly lapses is worse than no agreement — the customer still believes they are covered. The mechanics are in the service agreement and membership plans guide.
Batteries deserve a specific note, because they are the single most common reason a standby generator fails to start, and they are entirely predictable on a calendar. A company that tracks battery age per unit and proactively replaces them converts the most common emergency into a scheduled visit. That is the whole thesis of preventive service in one component.
| Call type | Clock | Intake must capture | Correct outcome |
|---|---|---|---|
| Existing unit, power out | Hours | Model, symptom, cranks or dead, fuel, controller display | Immediate dispatch or escalation, never a next-day callback |
| Existing unit, fault light | Days | Model, fault code, last exercise, agreement status | Booked service slot with history attached |
| Annual service due | Weeks | Nothing new — it is already on the cadence | Routed into a clustered maintenance day |
| Installation enquiry | Weeks | Panel, fuel, loads to carry, timeline, property type | Tier range on the call, booked load assessment |
| Orphaned unit, wants takeover | Days | Model, age, known history, why previous provider ended | Service visit plus agreement proposal |
Remote monitoring changes what "service" means
Modern standby units can report status remotely — exercise cycle results, fault conditions, whether the unit is in a ready state. That capability quietly changes the service model from reactive to notified.
The operational value is only realised if somebody acts on the alert, and that is where most implementations fail. An alert that lands in an inbox nobody reads is worse than no monitoring, because the company now technically knew.
The workable pattern is to treat a monitoring alert as an inbound service event: it creates a record, it is assigned, and it is either resolved or scheduled. A fault reported on a Tuesday afternoon becomes a Wednesday visit, and the customer hears from you before they notice anything. That is the strongest retention mechanic in the trade — a customer who gets a call saying "your generator reported a fault on its test cycle, we are coming Thursday" does not shop for another provider.
It also inverts the usual emergency ratio. Units under monitored agreements fail during outages far less often, because the failures were caught during the weekly exercise rather than during the event.
The post-storm wall
Then there are the weeks where everything above stops applying.
A major storm or freeze produces a call volume that has nothing to do with your normal week, and it arrives in a specific composition: existing customers whose units failed, existing customers wanting reassurance, and a very large wave of new prospects who have just spent two days without power and decided they are done with that.
Three things are true about that wave simultaneously. The emergency calls are the only time-critical ones. The reassurance calls are cheap to handle and hugely valuable to retention. And the prospect wave is the best-converting installation demand of the entire year, arriving at precisely the moment you have no capacity to talk to any of them.
The failure mode is universal: everything goes to voicemail, the emergencies are found hours late, and the prospect wave goes to whoever answered. Two weeks later the phones are quiet again and the surge is gone.
What actually helps is capacity that does not depend on your staff count. Every call answered and triaged, emergencies escalated immediately, reassurance calls handled at the front door with the unit's record in view, and the prospect wave captured with full intake and queued for assessment booking as capacity opens. The prospects do not need to be served this week — they need to be captured this week, because the decision they made during the outage fades within about a fortnight.
The general shape of this is in the seasonal call volume spikes guide, and the sharp-onset version — no warning, all at once — is in the storm and freeze call-surge playbook.
For companies that also rent portable units during outages, the fleet-and-utilisation side of that overlaps with the equipment rental operations guide, and the two demand curves peak in the same hour, which is worth planning for rather than discovering.
Permits, inspections and the project middle
The part of a generator installation that damages customer relationships is rarely the work. It is the middle — the three weeks between deposit and commissioning when the customer has paid, nothing visible is happening, and they do not know why.
The actual reasons are ordinary: equipment lead time, a permit in review, a gas utility scheduling a meter upgrade, an inspector's next available slot, an electrical inspection that has to precede the gas one. None of that is a problem. All of it becomes a problem when the customer cannot see it.
The operational answer is not faster permits; it is visible status. A job record with dated stages the customer can be told about, and a proactive touch at each transition — equipment ordered, permit submitted, permit approved, install scheduled, inspection booked, commissioned. Each is a thirty-second message that prevents a suspicious phone call.
The platform's AI outbound follow-up calls fit here precisely: touching your own existing customers about a stage change, a scheduled inspection or a pending approval, which is exactly the kind of proactive contact nobody has time to make manually and every customer remembers receiving. Whether that sits on your plan is detailed on the pricing page.
What to automate, what to keep human
Automate the triage question, the intake fields, the maintenance cadence, the unit history, the renewal, the stage notifications, the monitoring-alert-to-job-record path, the invoice and the review request. All of it is mechanical, all of it is currently inconsistent, and none of it requires a licence.
Keep human every technical judgement: the load calculation, whether a gas supply is adequate, whether a panel needs upgrading, whether an orphaned unit is worth taking on or is a liability, and whether a fault is a battery or a control board. Those need a technician and a manual, and a system that guessed would be wrong in the expensive direction.
The line is the same everywhere: if it needs a licence or a look, keep it. If it needs remembering, answering or repeating, automate it.
Where to start
Start with triage, because it is the cheapest change with the largest effect: one question at the front of every call that routes an emergency differently from a project. Everything else in this guide is downstream of getting that right.
Then get the unit record real — model, serial, fuel, install date, battery age, service history — because every fast service call depends on it, and every agreement renewal does too.
Then the agreement cadence and its renewal, then the stage notifications that make the project middle survivable, then storm-week capacity so the best installation demand of the year does not land in voicemail.
Plans, minutes and tier features are on the pricing page — three flat monthly options, no setup fee, month-to-month. To size it against your own install pipeline and service base, get in touch.
