After-hours maintenance calls: a simple playbook that lets your team sleep

Most after-hours maintenance calls aren't emergencies. A few are, and those few cost real money if they wait until morning. A burst pipe left for six hours turns a plumbing call into a ceiling, a floor and an insurance claim.

The goal of an after-hours process is simple: real emergencies reach a person within minutes, and everything else becomes a clean request your team picks up in the morning. The steps below work the same whether a person answers your phone at night or an AI voice agent for property management does.

Write down what counts as an emergency

If the rule only lives in your head, whoever answers at 2am will guess. Put it on one page. A typical list:

Call it an emergency	Usually routine (log it for the morning)
Active water leak, flooding, or water through a ceiling	Slow drip under a sink, running toilet
No heat when it's cold outside (set your own temperature line)	Thermostat questions, uneven heating
Gas smell (tell the caller to leave and call the gas company or 911 first)	Appliance not working (dishwasher, dryer)
Sewage backing up	Slow drain
No power to the whole unit	One outlet or light out
Exterior door or window that won't lock	Interior door or cabinet issues
Lockout, if your policy covers it after hours	Lost mailbox key

Two things make this list work. First, be specific about the edges. "No heat" in July is not the same call as "no heat" in January, so write down the outside temperature that makes it urgent. Second, decide what callers are told about their safety. For a gas smell, leave the unit and call the gas company or 911. For fire or anyone in danger, call 911. Leave repair steps, like finding the water shut-off, to your on-call person, who knows the building.

Collect the same details on every call

Half-heard voicemails are where requests go to die. Whoever answers should collect, every time:

1. Name and the best number to call back
2. Property address and unit
3. What's happening, in the tenant's own words
4. Whether it's getting worse (leaking now? spreading?)
5. Whether it affects other units, such as water coming from upstairs
6. Permission to enter if they're not home, plus pets or alarm codes the technician should know

Read the address and unit back before moving on. Street names and unit numbers are easy to mishear on a phone line, and a technician at the wrong door is a wasted night.

If you can't act on a request without calling the tenant back, it wasn't collected properly.

Decide exactly who gets woken up

An emergency needs a person, not a ticket. Keep an on-call rota with a first contact and a backup, and decide how they're reached: a transferred call, a text, or both. Then decide what happens if neither answers within, say, ten minutes. "We'll keep trying" isn't a plan. A named third person is.

For everything that isn't an emergency, the opposite rule applies: don't wake anyone up. Log it, tell the tenant when to expect a call, and let your on-call person sleep. Someone woken three times a week for a running toilet starts to dread the phone. That's the person you need sharp when a real leak comes in.

Close the loop with the tenant

The tenant should hang up knowing what happens next. Before the call ends, give them a short summary:

I've logged a leak under the kitchen sink at 12 Oak St, Unit 4. You've given permission to enter after 10am. Someone from our maintenance team will call you tomorrow morning.

That one summary heads off the second and third "just checking" calls. Put the same words in the log, so your team knows exactly what was promised. And only promise what your team can keep. If your coordinator can't call before noon, don't say 9am.

Hand off cleanly in the morning

By 9am your coordinator should have a list, not a voicemail box. Each item should show what was reported, how it was classified, and what was already done: who was called, whether they answered, and what the tenant was told. Urgent items should already be in motion.

Once a week, look back at the night's calls. If something was escalated that didn't need to be, or a real problem waited until morning, fix the emergency list. That list is what the person, or the agent, follows next time.

Where an AI voice agent fits

This playbook is the kind of work an AI voice agent does well. It asks every question on every call. It follows your emergency list the same way at 2am as at 2pm. And it doesn't get tired of the fifth dripping faucet of the night.

With Utter's maintenance agent, you choose what counts as an emergency: water leaks, no heat, gas smells, fire or smoke, and any others you add. You also choose what happens next: transfer the caller to your on-call person, log the request, alert your team, or all three. For everything else, it can walk the caller through a simple check first, then log the request and alert your team for the morning.

Whatever you use, write the rules down first. The answering is the easy part. The rules are what make it work. To see how any agent handles both kinds of call, start with the first two calls in 10 test calls to make before you buy an AI phone agent.