Which Property-Management Tasks Would You Automate First?

Gregory — I’d automate the transitions between tasks before I automated the tasks themselves.

That sounds semantic, but I think it’s the difference between useful automation and a pile of disconnected bots.

For example, “respond to rental inquiries” isn’t really one task. The actual workflow is:

Inquiry → qualification → scheduling → showing → follow-up → application → incomplete-application chase → screening → decision → lease

If you automate only the first step, you’ve saved a few emails.

If you automate the handoffs across the entire workflow, you’ve changed the economics of the operation.

So my order would probably be:

1. Lead response + pre-screening + showing scheduling

This is the obvious first target because speed matters enormously. A prospect who contacts you at 7:30 PM should not have to wait until someone opens an inbox the next morning.

The system should be able to answer the common questions, collect objective qualification information, schedule the showing, issue reminders, and follow up afterward.

But it should not decide who gets the property.

That boundary matters.

2. Maintenance intake and triage

Not necessarily “AI chooses the plumber.”

More like:

Tenant reports “sink leaking.”

The system immediately determines:

• active leak or slow drip?

• shutoff valve accessible?

• water reaching electrical?

• emergency or routine?

• which property/unit?

• photos/video available?

• existing warranty/vendor?

• tenant available when?

Now the human or vendor receives a structured incident instead of “my sink is broken, please call me.”

That’s a massive operational improvement.

3. Exception tracking

This is the part I think most automation conversations miss.

Property management doesn’t usually collapse because nobody knew how to send a rent reminder.

It collapses because:

• vendor never called back

• tenant never finished the application

• owner never approved the repair

• invoice never got matched

• maintenance ticket quietly sat open

• showing happened but nobody followed up

• lease expires in 43 days and nobody noticed

The highest-value system may be the one constantly asking:

“What was supposed to happen next, and did it?”

That’s different from a chatbot.

Dan mentioned building an AI agent themselves, and I think that’s increasingly realistic. But I’d be careful with the phrase “handles most of the tasks.”

I’d want the automation architecture to distinguish three categories:

Machine can execute.
Routine, reversible, rules-based.

Machine can recommend.
Requires judgment but can assemble the evidence.

Human must decide.
Applicant approval, sensitive tenant issues, legal/compliance questions, major expenditures, unusual exceptions.

Once those boundaries are explicit, automation becomes much safer and much more powerful.

So if I had to pick just one thing?

I wouldn’t automate rent reminders or FAQs first.

I’d automate the operational nervous system:

Capture every event, determine the next required action, execute the routine ones automatically, and relentlessly surface anything that falls outside the normal path.

Because the real enemy in property management isn’t repetitive work.

It’s dropped balls.