A property manager recently asked me an interesting question: how could AI help manage maintenance requests.
- They manage multiple residential buildings, each with more than 350 residents. Every day, their team deals with dozens of maintenance-related calls and emails.
- An elevator isn’t working.
- There’s a spill in a hallway.
- A resident notices water where it shouldn’t be.
- A tree or shrub has died.
- A light is out.
- Something is damaged in the parkade.
None of these requests is particularly complicated on its own.
The problem is what happens when you multiply them across several large buildings.
Every request has to be read, understood, categorized, prioritized, routed to the right person, acknowledged, followed up on and eventually closed.
And sometimes 15 residents are reporting the exact same broken elevator.
So the question I started with wasn’t:
Can AI answer maintenance emails?
Of course it can.
The much more interesting question is:
Can AI help manage the operational work created by those emails while still keeping people in control?
What is actually happening behind every maintenance request?
A resident might send one sentence:
“There’s water leaking from the ceiling by the elevators.”
But behind that one sentence is a workflow:
Receive → understand → identify location → determine urgency → check for duplicates → acknowledge → assign → approve → dispatch → follow up → update resident → close
The physical repair may take a contractor an hour.
But there can be a surprising amount of administrative work before and after that hour.
That is the part of the problem I became interested in.
Not replacing the plumber.
Not replacing the property manager.
Reducing the work around the work.
My first question: Does software already exist for this?
Before building anything, I would always look at what’s already on the market.
And there are several interesting platforms operating in this space.
Property Meld
Property Meld is particularly interesting because its focus is property maintenance rather than trying to become every piece of software a property manager uses.
Its platform handles maintenance communication, repair intake, troubleshooting, workflow and coordination between residents, property managers, technicians and vendors.
Its Core offering currently includes conversational AI repair intake, emergency screening, service issue creation and other maintenance workflow capabilities.
For an organization whose primary pain is maintenance operations, this would be one of the first platforms I would investigate.
Entrata
Entrata is a much broader property-management platform, but its Maintenance AI capabilities caught my attention.
Its software can interact conversationally with residents, help categorize and prioritize problems, gather information about the issue and reduce duplicate requests.
That last one matters more than it might initially appear.
If an elevator goes down at 8:00 AM, twenty residents could report it before 8:30.
You don’t have twenty maintenance problems.
You have one maintenance problem and twenty people who need communication.
Those are very different things.
AppFolio
AppFolio is another broader property-management platform with extensive maintenance capabilities.
Residents can submit maintenance requests and photos, while maintenance workflows can help with triage, vendor coordination and communication.
If an organization were also looking to consolidate other property-management functions, AppFolio would be worth evaluating rather than solving maintenance in isolation.
Building Engines
Building Engines takes a strong building-operations approach through its Prism platform.
It supports work orders, vendors, preventive maintenance, tenant communication and other facilities-management functions.
For larger or more operationally complex properties, where the problem extends beyond resident service requests into overall building operations, this becomes particularly interesting.
Zendesk
Zendesk might seem like the odd one out.
It isn’t property-management software.
It’s customer-service software.
But think about the original problem:
Dozens of incoming emails and calls need to be understood, organized, prioritized, routed, tracked and answered.
That is also a customer-service problem.
Zendesk already knows how to take communications from different channels, turn them into tickets, route them through workflows and maintain an audit trail.
The property-specific maintenance layer would need to be added or integrated, but I wouldn’t dismiss this approach.
So which one would I choose?
I wouldn’t choose yet.
And I definitely wouldn’t start building anything yet.
I’d first understand exactly what the property manager needs.
- Do they want a complete maintenance-management platform?
- Do they need technician scheduling?
- Do contractors need access to a portal?
- Do they need invoices and cost approvals?
- Do they already have a property-management system they like?
Or is their biggest problem simply this:
“We’re spending too much time reading maintenance requests, figuring out what they mean, figuring out who should handle them and keeping residents updated.”
Those are very different requirements.
And that’s where the build-versus-buy question starts getting interesting.
What if they only need part of the solution?
Suppose the organization already has accounting software.
- It already manages leases.
- It already has resident information.
- It already has contractor relationships.
- It already knows which elevator company services which building.
The missing piece may simply be an intelligent coordination layer.
Something that can:
- receive maintenance requests through email or a web page
- use AI to understand what the resident is reporting
- identify the building and location
- determine the likely category and urgency
- recognize when multiple residents are reporting the same problem
- acknowledge the resident automatically
- identify the appropriate contractor or employee
- draft the required communication
- place the work into a review queue
- let a property manager approve, change or reject the action
- contact the contractor by email or SMS after approval
- track the response
- automatically update affected residents as the issue progresses
That isn’t another property-management system.
It’s a much smaller application.
And importantly, AI doesn’t have to be given authority to spend money or dispatch contractors on its own.
AI can do most of the preparation.
A person can still make the decision.
That distinction matters
There’s a tendency in conversations about AI to jump directly to:
“How much can we automate?”
I’d approach it differently.
I’d ask:
“Which decisions should still belong to a person, and how much work can software remove before that person has to make the decision?”
Imagine a property manager opens a maintenance queue and sees:
HIGH PRIORITY
123 Main Street
Possible active water leak near P2 elevator lobby.
Six residents have reported the issue in the last 12 minutes.
No similar open incident was found.
Recommended contractor: ABC Plumbing.
And underneath:
- Approve Dispatch
- Assign Internally
- Change Contractor
- Request More Information
- Hold
That’s where I think AI gets much more practical.
It isn’t replacing the property manager.
It’s taking several minutes of administrative work and turning it into a decision that may take a few seconds.
Do that dozens of times every day across multiple properties and the savings begin to compound.
Build or buy?
That’s where I’m going next with this.
There are clearly capable commercial products already solving significant parts of the problem.
The next questions are:
- How much do they cost at scale?
- How much of their functionality would this particular property manager actually use?
- And if we built only the maintenance coordination layer ourselves, what would that cost compared with buying an existing platform?
That’s a much more interesting calculation than simply asking whether AI can answer an email.
Because sometimes the best solution is to buy.
Sometimes it’s to integrate.
And occasionally, when the problem is sufficiently narrow and the volume sufficiently large, building exactly what you need may make sense too.
This Is Part 1 of a 5-Part Series
This article is the starting point for a five-part series on how AI could practically be used to solve this kind of operational problem.
Over the next four parts, I’ll break down:
- the existing software options and where they fit
- what a purpose-built maintenance coordination system could look like
- the high-level product architecture, workflows and AI components needed to build it
- the economics of building versus buying
- where human approval should remain in the process, even when AI is doing much of the work
The goal is not just to talk about what AI could do.
I want to get specific enough that a property-management company, product team or developer could look at the proposed workflow and understand how a system like this could actually be built.
Next: What would the custom AI-powered maintenance system look like, from the first resident email through contractor dispatch and resolution?
About The Author
