Book a call

Case studyExample business, real tools, estimated numbers

Seven AI agents run a property management company's night

In this case study, seven AI agents run the night of a small property management company. At 2 AM a tenant calls with water coming through her ceiling. The agents keep her safe, walk her to the shutoff valve, send the on-call plumber, text the unit upstairs and hold the $640 repair for the owner, whose limit is $500. At 7 AM the property manager gets one message with the two things that need her.

The video loads from YouTube only when you press play.

Business
Northline, a small property management company, building Maple Court
Size
A property manager and an on-call plumber, with owners who set repair limits
The night
11 PM to 7 AM
Agents
Seven: Otto, Front Desk, Maintenance, Residents, Leasing, Books and Owners
Tools
AppFolio QuickBooks Claude
Build time (estimate)
About three weeks, then the AI officer plan
40sec
to answer an 11 PM tour request (from the next morning)
One night
1message
for the manager at 7 AM (from a night of voicemails)
One night
0
repairs over the owner's limit approved without the owner
By design
Cover of the guide: The AI Night Shift

Free guide · PDF · 79 pages

The AI Night Shift

Implementation guide for property managers: the seven agents with their build specs, the emergency rules, the rollout and the templates.

View the guide Download PDF

The situation

Maintenance is where owners judge a manager. In a 2026 Buildium survey of 300 rental owners, maintenance was the top stressor for 46% of them, 59% rated it the most valuable service a manager provides, and 84% wanted to approve large repairs before the work is done.

Emergencies do not wait for office hours. Water through a ceiling at 2 AM needs a calm voice, the right steps in the right order, a plumber, a word to the unit upstairs, and a decision about money from someone who is asleep.

In this case study, the call comes from unit 4B at Maple Court. The water is dripping right by the kitchen light.

How the night used to go

At a small company, an after-hours call reaches an answering service that takes a message, or wakes the manager. Someone finds a plumber, someone else remembers the unit upstairs, and the owner hears about the bill the next day, often after the work is done.

Meanwhile, the tour request that came in at 11 PM waits until morning. By then the prospect has booked a tour somewhere else.

The agents

Seven agents, each with one job and a written rulebook, all working inside the company's own tools. Otto, the operations chief, routes every call, text and email.

OttoLead

Routes every call, text and email to the right specialist, keeps the log, and sends the manager one message at 7 AM.

Never makes a money or policy decision.
Front Desk

Answers every call, says it is the company's AI assistant, keeps the resident safe first, then follows the emergency policy word for word.

Never improvises outside the written policy.
Maintenance

Opens the work order in AppFolio, works out the likely source (the unit upstairs) and dispatches the on-call plumber.

Never approves a cost above the owner's limit.
Residents

Texts the tenant upstairs before anyone knocks, and keeps every resident updated.

Never shares one tenant's details with another.
Leasing

Answers tour requests in seconds, day or night, and books the tour.

Never screens applicants or decides who gets a lease.
Books

Matches the plumber's invoice to the work order and the owner's approval.

Never pays a bill on its own.
Owners

Asks the owner to approve any repair over the limit, with photos and the plumber's note.

Never approves for the owner.

Step through the night

Press play, or tap any time, to follow the night from the tour request to the 7 AM message.

One night
11:00 PM
Leasing

A prospect asks for a tour on the company's site. The answer comes back in 40 seconds, with three open times.

1 / 8

The stack

No new logins. The agents work inside what the company already pays for.

AppFolio, units and work orders QuickBooks, the books Claude, the agents
  • AppFolio holds the units, the residents, the owners and the work orders. Maintenance opens the order there, Owners reads the approval limit there.
  • QuickBooks keeps the books. Books matches each invoice to its work order and approval.
  • The phone line forwards after-hours calls and texts to Front Desk, which always says it is an AI.
  • The emergency policy is the company's own document. Front Desk reads it, word for word.

Before and after

MomentBefore (typical)With agents
Call at 2 AMAnswering service takes a messageAnswered, safety first, valve turned in a minute
The plumberFound by whoever is awakeDispatched from the work order at 2:03 AM
The unit upstairsFinds out when someone knocksTexted before anyone knocks
A repair over the limitOwner hears about it afterwardsOwner approves first, with photos
Tour request at 11 PMAnswered the next morningAnswered in 40 seconds
The manager's morningVoicemails to sortOne message, two decisions
 Before: a typical management company, estimated. With agents: the night in this case study.

The value, and how we estimated it

Counted in hours and outcomes, with each assumption written out so you can redo the math with your own management company's numbers.

ValueEstimateHow we estimated it
Manager's nights backmost of themAfter-hours calls follow the written emergency policy; the manager is woken only when the policy says so.
Tour requests answered after hoursall of themLeasing answers every request in seconds instead of the next morning.
Repairs over the owner's limitevery one approved firstOwners sends the request with photos before the work goes further than stopping the damage.
Invoices matchedthe same morningBooks ties each invoice to its work order and approval.
 Estimates, with every assumption shown.

How a build like this would run

Estimated for a company of two to ten people on AppFolio and QuickBooks.

  1. Week 1

    Map

    Read the emergency policy and the owner agreements, listen to a month of after-hours calls, list the on-call vendors and each owner's repair limit. Agree on the result, for example: every after-hours call answered, and every repair over the limit approved first.

  2. Weeks 2 to 3

    Build

    Front Desk and Maintenance first, on the after-hours line. Then Residents, Owners and Books. Leasing last, with fair housing rules written in.

  3. Day 30

    Prove

    Measured from the phone log and AppFolio: calls answered, time to dispatch, approvals before work.

  4. Every month

    Stay

    A report, policy updates when an owner changes a limit, and one improvement a month.

What stays with the people

The agents answer, dispatch and document. People approve money and make the judgment calls.

Safety first, policy word for wordFront Desk follows the company's written emergency policy and never improvises around it.
Gas is a routing ruleA gas smell means leave the building, no switches, call 911 or the gas company from outside, then alert the on-call manager.
Owners approve big repairsAnything above the owner's limit waits for the owner, except stopping active damage.
Fair housingLeasing gives every prospect the same answers and never screens or decides who gets a lease.
The AI says it is AIOn every call and text, and it texts only people who said yes.

What's in the guide

The full implementation guide behind this case study, 79 pages, free to read and to download.

  1. The business reality: how small management companies run, the tools they use and where the hours go, with sources.
  2. One night at Northline, minute by minute.
  3. The seven agents: a job card and a build spec for each.
  4. The rules: emergency policy, gas, owner approval limits, fair housing, privacy and texts.
  5. Build and run it: the stack, the data flow, the 30-60-90 day rollout and the numbers to track.
  6. Templates and appendices: the Claude skill and its files, scripts, a glossary and every source.

View the guideDownload PDF

Where the numbers come from

The tools and rules are real. The company, the people, the building and the amounts are invented for this case study, and the before and after are estimates.

Read the video transcript

Tenant: Hi, I'm so sorry, I know it's two in the morning, but there's water coming through my ceiling!

Front Desk (AI): You did the right thing calling. First, are you safe? Is the water near any lights or outlets?

Tenant: It's dripping right by the kitchen light.

Front Desk (AI): Okay. Don't touch that switch. There's a valve under your kitchen sink. Turn it to the right for me.

Tenant: Okay... okay. It's slowing down.

Front Desk (AI): Perfect. A plumber is on his way. I'm texting you his name and photo right now.

Narrator: That's not a person picking up. It's one of seven AI agents running an entire property management company, all night. And the blueprint is free.

Narrator: At the top sits Otto, the operations chief. Every call, text and email lands here first, and Otto hands the work to six specialists, inside the tools you already use. AppFolio, QuickBooks, your phone line. No new logins.

Narrator: Front Desk keeps the resident safe first, then follows your emergency policy, word for word. Maintenance opens the work order, and figures out the water is probably coming from the unit upstairs.

Maintenance (AI): Hi Mike, it's Northline. Active leak at Maple Court, unit 4B, probably coming from 5B upstairs. Can you take it?

Mike: Yeah... yeah, I got it. I'll be there in thirty.

Narrator: The Residents agent sends a text to the tenant upstairs before anyone knocks.

Narrator: Mike stops the leak. But the repair is six hundred and forty dollars, over the owner's five hundred dollar limit. And eighty-four percent of owners want to approve big repairs first. So the repair waits for a human. At seven, the owner taps approve, and Books matches the invoice.

Narrator: Meanwhile, Leasing answered an eleven P.M. tour request in forty seconds.

Narrator: And at seven A.M., the property manager gets one message: what happened overnight, and the two things that need her.

Narrator: These aren't seven separate chatbots. They hand work to each other, like a team that never sleeps. AI does the work. You make the decisions. Comment BLUEPRINT, and I'll send you the whole system.

For property managers

Want this for your company?

In 45 minutes we look at how an after-hours call reaches you today, and tell you if an agent can take it inside the tools you already use.

Get a free 45-minute map

More case studies

An insurance agency's busiest night7 agents · Insurance agencyA roofing company's storm night7 agents · Roofing company