Build it with AI
Build a focused property workflow for units, requests, vendors, and operating status.
Create a system around property records, maintenance intake, vendor coordination, inspections, and owner reporting.
Introduction
What a property management system has to hold.
Most teams end up with a property management system the same way: tenancy details in one system, maintenance in email, and compliance dates in a spreadsheet. Handling requests as they arrive works up to the point where the number of properties exceeds what one person can hold. Past that, a request received by phone on Tuesday is indistinguishable from one that was never made.
Compliance dates sit in a spreadsheet nobody owns, and the consequence of missing one is legal rather than administrative. There is no record of a request until somebody acts on it, no history against the property that outlives a tenancy, and no visibility of which vendor jobs are outstanding.
What follows covers building a property management system: the records it holds (properties, tenancies, rent, maintenance requests, compliance certificates, and renewal dates), the systems it reads (Google Calendar and Gmail), and what it does not fix.
The problem
Maintenance requests arriving through four channels and one memory.
Property management platforms are comprehensive and priced for portfolios larger than most operators run. Spreadsheets track the properties and nothing about the requests, which is where the work and the complaints both come from.
The records are properties, tenancies, rent, maintenance requests, compliance certificates, and renewal dates, and the authoritative copy of most of them already lives in Google Calendar or Gmail. The tenant says they reported it three weeks ago, the log has nothing, and the report was a text message to a mobile.
The cost is not the inconvenience: a certificate expires on an occupied property.
You're likely here because
- A tenant can report something and have no record exist
- Compliance dates sit in a spreadsheet nobody owns, and the consequence of missing one is legal rather than administrative.
- When it is wrong, a certificate expires on an occupied property
What gets built
Launch builds it, Grow operates it.
Built in Launch
- • Property records
- • Maintenance queue
- • Vendor tracker
Operated through Grow
- • Notifications
- • Scheduling
- • Follow-up
Systems it reads
- • Google Calendar
- • Gmail
- • Google Drive
The record model
What a property record has to carry.
- Request captured at first contact
- From phone, message, email, or form. A request that exists in a conversation and nowhere else is where every dispute originates.
- Property history across tenancies
- So a recurring fault is visible as recurring rather than as three separate first-time reports by three different tenants.
- Urgency and habitability classification
- Escalated by rule rather than by triage judgement, given the legal exposure attached to delay on habitability issues.
- Vendor work order with scoped access
- The contractor sees the job, not the tenancy or owner detail behind it.
- Tenant-visible status
- Because most chasing contact is a request for confirmation that the report was received, not for a completion date.
- Inspection record with date and outcome
- Held against the property, since inspection obligations recur on a cycle and are evidenced rather than asserted.
- Owner statement source
- Generated from the same records as the maintenance log, so the two cannot disagree and produce a dispute.
How it runs
From scattered requests to tracked property state.
Step 01
Describe what a property management system has to do
Model the request as the core record, attached to a property and outliving any single tenancy. Property history is the asset; the current request is the work.
Step 02
Connect the systems of record
Email and messaging are where requests actually arrive, the calendar holds vendor appointments, and the document store holds inspection records. Capturing from the real channels is what makes the log complete.
Step 03
Build the operating surface
Property records with history, a maintenance queue with status and vendor assignment, inspection scheduling, and owner reporting.
Step 04
Start narrow
Request capture from every channel into one queue with a status a tenant can see. Capture before workflow — an incomplete log makes every downstream number wrong.
Step 05
Route the exceptions
A request that reads as urgent or as a habitability issue escalates immediately rather than queueing, since the consequences of delay there are not proportionate.
Step 06
Measure compliance items actioned before expiry rather than after
Measure time from request to resolution and the share of requests captured with a record at first contact. The second determines whether the first means anything.
Implementation path
Building property operations around the request.
- 01
Capture requests from every channel people actually use, including phone and message. A queue fed only by a portal reports excellent performance on the requests it can see.
- 02
Baseline time to resolution and the share of requests with a record at first contact. The gap between reported and logged is usually the real problem.
- 03
Attach history to the property rather than the tenancy, so a recurring fault is visible as recurring rather than as three separate first-time reports.
- 04
Give tenants status visibility early. Most chasing contact is a request for reassurance that the report exists at all.
- 05
Build the narrowest useful version first: compliance and renewal dates surfaced with enough lead time to act, against a named owner.
- 06
Capturing requests from every channel into one queue is the first two weeks and is the prerequisite — a queue fed only by a portal reports excellent performance on a minority of requests. Tenant-visible status is a further week and removes a large share of inbound contact. Escalation rules for habitability should be reviewed against local requirements rather than set by judgement.
- 07
Once capture and status are working, add vendor work orders with scoped access, then owner statements generated from the same records. Inspection scheduling follows the obligations cycle.
Controls
Controls that matter.
Control 01
Urgent and habitability-related requests escalated immediately by rule rather than by triage judgement, given the legal exposure attached to delay
Control 02
Property history retained across tenancies, since a recurring defect is only visible over a period longer than one tenant
Control 03
Vendor access scoped to the job, so a contractor sees the work order and not the tenancy or owner detail behind it
Examples
Three requests that stop getting lost.
The report that was a text message
Capture from messaging and email into the same queue closes the gap where a request exists in a conversation and nowhere else, which is where disputes originate.
The boiler that failed three times
History attached to the property rather than the tenancy turns three separate incidents into one visible pattern, which changes the decision from repair to replace.
The tenant chasing for an update
Status visibility removes most chasing contact, which is largely a request for confirmation that the report was received rather than for a completion date.
How it goes wrong
Three ways property operations lose track.
Reporting is portal-only, and the requests that become disputes arrived by text message.
Accept requests however tenants already report them. A clean queue that omits the contentious requests is worse than a messy one that contains them.
History is attached to the tenancy, so a boiler that has failed three times looks like three first reports.
Attach history to the property. Recurring faults are only visible over a period longer than one tenancy, and that is exactly the decision point between repair and replace.
Habitability issues are triaged by whoever is on the queue that day.
Escalate by rule, reviewed against local obligations. Delay here carries consequences that are not proportionate to the effort of a classification rule.
Limitations and considerations
What tracking will not fix about the work itself.
- Tracking a request does not repair anything. Where the constraint is vendor availability, better visibility surfaces the delay earlier and does not shorten it.
- Habitability, safety, and notice obligations vary by jurisdiction and carry real consequences for delay. Escalation rules should be reviewed against local requirements rather than set by general judgement.
- Vendor coordination remains manual to the extent that vendors are not on any shared system. Expect the workflow to end at a phone call for a meaningful share of jobs.
- Below the point where one person can hold the outstanding requests — typically twenty to fifty units depending on age and tenant mix — this is overhead. If the delay is vendor availability, better visibility surfaces it earlier and does not shorten it.
- Connector coverage varies: Google Calendar, Gmail, Google Drive are representative rather than guaranteed, and the fields exposed depend on your workspace permissions.
FAQ
Build a property management system with AI: common questions.
How many properties before this is worth building?
The threshold is where one person can no longer hold the outstanding requests in their head, which is typically somewhere between twenty and fifty units depending on age and tenant mix. The signal is disputes about whether something was reported.
How should tenants report issues?
However they already do — phone, message, email, or a form. A portal-only channel produces a clean queue that omits the requests most likely to become disputes, which defeats the purpose.
Can it dispatch vendors automatically?
It can create and route work orders with the property context attached. Fully automatic dispatch depends on vendors being on a shared system, which most small operators cannot assume.
What about owner reporting?
Generate it from the same records rather than assembling it separately. Owner statements that disagree with the maintenance log are a recurring source of disputes and are entirely avoidable.
What should the first version contain?
Compliance and renewal dates surfaced with enough lead time to act, against a named owner. Everything else waits until that one is genuinely used.
How will we know whether it worked?
Measure compliance items actioned before expiry rather than after against the baseline taken before anything changed.
Start with ARIA
Ask ARIA to build it.
Describe the website, application, workflow, or operating surface you need. ARIA plans, connects, builds, tests, and keeps refining it — inside the permissions you set.
- ARIA acts only through the systems and permissions you connect.
- Connections use scoped credentials you can change or revoke.
- Actions are recorded, and consequential ones can require approval.
Start here
Build a property management system around the process you actually run.
Capture requests from every channel first, attach history to the property rather than the tenancy, and escalate habitability issues by rule.