A surge is not a busy season
Every service trade has a busy season, and every owner knows roughly when it starts. That is a staffing problem, and it has a staffing answer: hire earlier, extend hours, put the part-time office person on full days from May through September. It is planned, and the guide to planning it is the seasonal call volume playbook.
A surge is a different animal entirely. A surge is a hard freeze that bursts pipes across four counties overnight, a hailstorm that lands in a twenty-minute window and puts holes in three thousand roofs, or a regional power event that takes down every commercial refrigeration compressor in a shopping district at once. It is not on a calendar, it does not build gradually, and by the time you know it is happening your phone has already been ringing for an hour. As of August 2026, the single sharpest difference between service businesses that come out of a surge with a full backlog and those that come out with a bad review cycle is not truck count or crew size. It is what happened on the phone during the four hours when everyone called at once.
This playbook is about that. It covers the queue math that decides how many callers you keep, why concurrent answering removes the queue instead of shortening it, how to triage true emergencies from work that can wait, how to book honestly when you are already sold out, how to work the callback list after, and what a surge week actually costs in AI call minutes on each Run with Jarvis plan. The math is all built from published plan numbers so you can rerun it against your own volume.
The queue math nobody runs until it is too late
Start with a number most owners have never calculated: your maximum answerable calls per hour.
Take one competent person answering the phone. A real service intake — greeting, understanding the problem, getting the address, qualifying urgency, quoting if you quote, booking, confirming — runs about five minutes. Some run three, some run eight, five is a fair working average. That person can complete twelve calls in an hour if they never take a breath, never step away, and never spend two minutes on a customer who is upset. Call it ten in practice.
Now put a hard freeze on top of it. A regional freeze event does not raise your call volume by thirty percent. It raises it by a factor, and the factor lands in a window. Two hundred calls arriving in one hour against a ten-call-per-hour ceiling means one hundred and ninety of those callers get something other than a person: a busy signal, a ring-out, a voicemail greeting, or four minutes of hold music.
Here is the part that costs the money. Those callers do not queue politely and wait for you. A homeowner standing in an inch of water on their kitchen floor at 6 a.m. is not evaluating vendors — they are dialing until somebody picks up. The dominant behavior in a surge is serial dialing, and the reward goes entirely to whoever answers first. Every second in that behavior is decisive, which is the same dynamic that drives ordinary lead response and is unpacked in the speed-to-lead guide. A surge just compresses it. Normally you lose a lead to a slow callback. In a surge you lose it to the ring itself.
Adding humans does not fix this, and it is worth being blunt about why. Suppose you could conjure three extra trained people the moment the freeze hit. Your ceiling moves from ten calls an hour to forty. Against two hundred, you still miss one hundred and sixty. You have quadrupled your payroll to recover twenty percent of a spike that lasts six hours. There is no staffing level that solves a surge, because a surge is defined by exceeding whatever staffing level you chose.
The second cost is invisible and larger. Voicemails from a surge are worth almost nothing by the time you return them. If two hundred people call between 6 and 10 a.m. and you start returning messages at 2 p.m., most of those people have already been served by somebody else. You spend a full afternoon of labor making calls that mostly end in "we got someone, thanks." You paid twice: once for the lost job, once for the callback that discovered it was lost.
Why concurrent answering removes the queue instead of shortening it
The structural fix is not a faster human or a bigger phone tree. It is concurrency.
An AI receptionist does not have a queue in the way a person does, because it does not handle calls one at a time. It holds simultaneous conversations. Six callers, forty callers, two hundred callers — each one gets answered on the first ring and gets a complete intake, in English or Spanish, at 6 a.m. on a Sunday during an ice event exactly as reliably as at 2 p.m. on a Tuesday. The concept of hold time stops applying, because nobody is waiting for a resource that is busy with someone else.
That single property is what converts a surge from a loss event into a backlog event. A backlog is an asset. Missed calls are not.
It also changes what "capacity" means for your business. Your trucks still have a hard limit — a surge does not create technicians. But your intake capacity becomes effectively unbounded, which decouples two things that were always fused when a human answered the phone. You can capture one hundred percent of demand while serving only the fraction of it you can physically staff. Every caller you cannot serve today becomes a named, qualified, contactable record instead of an anonymous missed call. That distinction is the entire premise of the rest of this playbook, and it is the practical difference described in what an AI employee actually is for service businesses.
The other half is that intake quality does not collapse under load. A human on their ninetieth call of a freeze morning is exhausted and starts abbreviating: they get a name and a number and skip the address details, the access notes, the equipment specifics. An automated intake asks the same questions in the same order on call two hundred as on call one. The records you work from at 4 p.m. are as complete as the ones from 6 a.m., which matters enormously for the triage and follow-up steps that come next.
Triage is the surge skill that matters most
In normal weeks, triage is a nicety. In a surge it is the whole game, because demand exceeds capacity and something has to decide the order.
Not every call in a freeze event is an emergency, even though every caller believes theirs is. Inside a single storm morning you will get an active burst pipe spraying water into a finished basement, a house with no heat and an infant in it, a homeowner whose outdoor faucet is dripping and wants it looked at eventually, someone asking whether their system needs winterizing, and three people calling to ask whether you are even open. Those five calls deserve five different outcomes, and only the first two justify a truck today.
Structured intake sorts them at the door. The AI runs your triage questions on every call — is water actively running, is there heat in the building, is anyone vulnerable in the home, is the property occupied, is there commercial inventory at risk — and the answers determine what happens next. That is not a judgment call the software invents. It is your own dispatch policy, applied consistently to two hundred calls instead of inconsistently to twelve.
Three things fall out of good triage that owners consistently underestimate.
You dispatch in the right order. Without triage, your first-come-first-served list sends a truck to a dripping faucet at 7 a.m. while a burst supply line waits until noon. That is not just a service failure, it is a margin failure, because the emergency job carries the higher ticket and the higher urgency premium.
You stop rolling trucks on work you should not take. A surge is when out-of-area calls spike hardest, because desperate people widen their search radius and call anyone who answers. A ninety-minute drive during a freeze week is one of the most expensive decisions you can make, since that truck could have completed two local emergencies in the same window. Your service radius should get tighter during a surge, not looser, and the drive-time economics behind that are laid out in the service radius and drive-time pricing guide.
You spot the work that is not yours. Restoration-adjacent surges are the clearest case. A caller with three inches of standing water needs mitigation before they need whatever you sell, and knowing that at intake lets you route them properly instead of sending a technician who cannot help. The overlap between emergency answering and water intrusion work is covered in the water damage restoration answering guide.
Capacity-honest booking, or how not to make forty promises you cannot keep
Here is where good surge operations separate from bad ones, and it has nothing to do with technology.
The temptation during a spike is to book everybody. The phone is producing more demand than you have seen in a year, and saying yes feels like winning. So the calendar fills with same-day and next-day promises, and by Wednesday you are running four hours behind on every appointment, techs are calling customers from the truck to push them, and the reviews arriving the following week say "they never showed up."
That is a self-inflicted wound. You converted a windfall into a reputation cost.
Capacity-honest booking means the calendar is the constraint and the intake respects it. The system holds real availability, so when today and tomorrow are genuinely full, the AI does not offer today and tomorrow. It offers Thursday. And the script matters: a caller told plainly "we are running about three days behind because of the freeze, I can put you down for Thursday morning and text you if something opens sooner" overwhelmingly accepts that, because they have already called four companies and heard nothing at all. Honesty is competitive in a surge specifically because everyone else is either not answering or overpromising.
Three rules hold up under pressure:
Never quote a window you have not staffed. An arrival window you miss costs more than the wider window you were afraid to say out loud.
Communicate the backlog proactively. Automatic arrival texts and ETA updates are worth more during a surge than at any other time, because the alternative is every one of your scheduled customers calling to ask where the tech is — which adds inbound volume on top of the surge you are already absorbing. "Where is my tech" calls are pure cost and they multiply exactly when you can least afford them.
Separate the emergency queue from the schedulable queue. True emergencies get dispatched, not booked. Everything else gets a real slot on a real calendar. Blending them means an emergency waits behind a routine visit, which is the worst possible outcome for both customers.
The judgment about which calls justify same-day dispatch after hours, and how to price that, is the same one covered in the after-hours calls playbook — a surge simply raises the volume without changing the logic.
The surge callback list is the actual asset
When the spike subsides, you are left with something most companies never had before: a complete list of people who wanted to buy from you and did not get served.
That list has three groups in it. People you booked for later in the week and need to confirm. People who called, got a full intake, and were told you could not reach them in their timeframe. And people who were captured but never scheduled at all, because they were on the edge of your radius or their issue was outside your scope on that particular day.
Historically that list did not exist. Surge overflow was voicemail, and voicemail was a shoebox nobody had time to empty. Now it is structured customer records with the problem described, the address captured, and the urgency scored.
Working it is where the platform's AI outbound follow-up calls do the most obvious work of the whole year. This is follow-up on your own captured leads — people who called you, not cold prospecting — and it is exactly the kind of volume no human office can absorb in the same week it is also digging out of a backlog. Two hundred outbound follow-up calls is a week of a person's life; it is an afternoon for the system. The general mechanics are in the AI outbound follow-up guide.
The 72 hours after the surge is where the margin lives
The spike gets all the attention and produces a minority of the revenue. Understand the shape of the demand curve.
During the spike itself, you are capacity-limited. You physically cannot do more work, so your revenue on those days is capped by truck count and daylight, no matter how well the phone is answered. The reason to answer the phone perfectly during the spike is not that it lets you do more work that day. It is that it records the demand you could not serve.
Then the curve flattens. Two or three days after a freeze, the panic ends. Your competitors are still swamped, still not answering, still working through a paper pile of voicemails. The people who could not get served during the peak are now shopping calmly instead of frantically — and they are still broken. That is the window where your captured list turns into booked work at normal, non-chaotic scheduling, with your trucks running efficient routes instead of emergency crisscrossing.
Practically, that means three moves in the seventy-two hours after the peak:
Contact every unserved caller. Not the ones who booked — the ones who did not. Lead with the acknowledgment that you could not reach them during the storm and that you have availability now.
Reconfirm everything on the calendar. Surge-week bookings have an elevated cancellation rate because some of those customers found someone else in the interim. Confirming turns a ghost appointment into either a kept job or a freed slot you can sell to somebody on the callback list.
Follow up on every quote that went quiet. Storm-driven work produces a lot of "let me talk to my spouse" and "let me check with insurance." Those stall out permanently unless something touches them again.
None of the three is complicated. All three are things that simply do not happen when the office is buried, which is precisely why automating them is worth more than any other single thing on this list.
What a surge week actually costs in AI minutes
Because plans include a minute pool with a flat published overage rate, you can price a surge before it happens instead of being surprised by it.
Build the scenario. Assume a normal month of 90 answered calls at a 4-minute average, which is 360 minutes. Surge calls run shorter than normal intakes — they are triage-heavy and the caller is in a hurry — so use 3 minutes. A moderate surge week adds 300 calls, or 900 minutes. A severe one adds 600 calls, or 1,800 minutes.
| Month scenario | Total AI minutes | Core all-in | Pro all-in | Elite all-in |
|---|---|---|---|---|
| Normal month, no surge | 360 | $500.00 | $750.00 | $1,200.00 |
| Moderate surge week added | 1,260 | $842.00 | $854.00 | $1,200.00 |
| Severe surge week added | 2,160 | $1,247.00 | $1,214.00 | $1,200.00 |
Core is $500 a month with 500 minutes and $0.45 per minute after; Pro is $750 with 1,000 minutes at $0.40; Elite is $1,200 with 2,500 minutes at $0.35. All three are month-to-month with zero setup fees and unlimited users, and current numbers live on the pricing page.
Read the table carefully, because it says something counterintuitive. At a moderate surge, staying on Core and eating 760 overage minutes costs $842 — twelve dollars less than Pro would have cost for the same month. One surge does not justify a permanent tier move. It justifies paying the overage and moving on.
At a severe surge the ordering flips: 2,160 minutes costs $1,247 on Core, $1,214 on Pro, and exactly $1,200 on Elite, because Elite's 2,500-minute pool swallows the whole event with no overage at all. If your region reliably produces a severe event more than once or twice a year, that changes the calculation.
The decision rule is the same one that governs any tier move: if your all-in cost lands near the next tier's sticker price for two consecutive months, upgrade, because you get more included minutes and a whole additional system for money you are already spending. A single freeze week is a cost of the storm. A pattern is a signal. Month-to-month billing is what makes it safe to treat them differently — you are never locked into a tier you bought for one bad week.
Build the surge kit before the sky does anything
Everything above works only if it is already configured when the event starts. Nobody sets up call handling during a hailstorm. A short pre-season pass:
Write your triage questions down and make sure the intake asks them — the specific field that decides emergency versus schedulable for your trade, not a generic urgency question. Set your genuine daily capacity in the calendar so the booking layer cannot oversell your trucks. Decide your surge radius in advance and stick to it when the phone is screaming. Confirm arrival texts and ETA updates are switched on, because they suppress the second wave of inbound volume. And decide now what your after-hours and emergency pricing is, so the intake can state it consistently rather than a tired dispatcher improvising at 3 a.m.
For general small business continuity planning, the Small Business Administration publishes preparedness material at sba.gov worth reviewing alongside your operational plan. The operational half — the phone — is the part nobody writes down, and it is the part that decides how much of the surge you keep.
The short version
A surge is arithmetic. One receptionist at five minutes a call handles about twelve an hour; a freeze morning delivers two hundred. The gap is not a work-ethic problem and no realistic amount of hiring closes it. Concurrent AI answering removes the queue entirely, which converts an unrecoverable loss into a captured backlog.
From there it is discipline: triage so the right truck goes to the right emergency first, book honestly so you do not spend the following week apologizing, tighten your radius rather than widening it, and treat the unserved caller list as the asset it is. Then work that list hard in the seventy-two hours after the peak, when demand is still real and your competitors are still buried. That window, not the spike, is where the margin sits.
If you want to walk through what your own surge math looks like against your call volume and average ticket, get in touch or compare the current minute pools on the pricing page.
