Built world
The office is not understaffed because the buildings got taller. It is understaffed because every request still needs a person to interpret it.
Work orders, renewals, vendor dispatch, and resident messages arrive in every channel. Someone reads, classifies, enters, chases, and updates. That is the job, and it grows linearly with units.
How I see it
Property management is a queue of incomplete requests
Property management looks like hospitality plus facilities plus billing. The unit of work is a completed request: a closed work order, a signed renewal, a resolved resident issue. AI is useful when it classifies, assembles, dispatches, and writes back.
The design error is a chatbot that answers FAQs while the work order still sits in an inbox. Residents will use any channel. The operating path has to land in one queue with a complete packet.
Human control points belong on spend, habitability, and anything that creates legal or safety exposure. Routine classification and status updates should not wait for a person who is already on the phone.
Start with one request type that already has volume. Measure time-to-complete and reopen rate. Do not launch a smart-building program.
Common mistakes
What teams usually get wrong.
Standing up a resident bot and leaving the work-order path alone
You added a channel. You did not add capacity.
Treating every property as a different company
Local habit is real. The request types are usually the same. Pick one portfolio slice and one request type.
Never writing back to the property system
If the bot talks and the PMS is empty, the office still has to type.
A useful diagnostic
Five questions before you fund the work.
Which request type consumes the most office hours?
That is the first workflow, not a general assistant.How many channels can the same request enter?
If the answer is 'all of them,' intake is part of the redesign.What does a delayed work order cost in churn, overtime, or vendor premium?
If you cannot estimate it, you cannot rank the bet.Which actions require a manager because of spend or safety?
Those stay human. Classification should not.Will site teams use the new path when the afternoon gets ugly?
If the old inbox is easier, the old inbox wins.
Economic model
Office loadunits × requests per unit × minutes of interpret-enter-chase time × loaded office cost
A conservative slice of that number is the first-year case. If it is small, the request type is wrong.
Three credible paths
How far should you go?
Do not force one solution. Choose the path the economics, the risk, and the organization can support.
Improve the work you already have
Keep the process mostly intact and use AI on the bottlenecks that create delay, rework, or follow-up.
The workflow is already sound and a few steps create most of the friction.
Gains are usually incremental. The operating economics do not change much.
Redesign the workflow around AI
Question every handoff, queue, and duplicate step, then rebuild the process around what AI can now do.
The process grew over years and coordination now costs more than the work itself.
Requires process change, clearer ownership, and a willingness to retire old steps.
Put an agent on a high-value outcome
Give an agent responsibility for one valuable result across systems, with humans at the control points that require judgment, authority, or risk acceptance.
The workflow is high-value, variable, multi-step, and worth engineering for production.
Needs stronger architecture, evaluation, controls, and monitoring. This is not a prompt project.
When this is the wrong next step
Do not start with a sector program.
- The company wants a consumer app for the board, not a closed work-order path.
- Each community manager will keep a private process.
- The PMS cannot be written to, and nobody will admit that is the constraint.
- Volume is a handful of luxury buildings with relationship-only service.
A useful next step
Bring one expensive workflow from this industry.
We guide the implementation, go deep on the technical path, and stay hands-on through operations — or tell you when a simpler answer is better.
Discuss an AI opportunity