CentriWebAI for property management is most useful on the three jobs that fill a property manager’s inbox: triaging maintenance requests, reminding tenants about rent, and turning inspection notes into reports. It should draft and sort, while a licensed person makes decisions that carry legal weight such as terminating a tenancy, authorising costly repairs or deciding who is responsible.
This article is for principals and property managers in Australia. Tenancy law is made by each state and territory, so the examples here use New South Wales and you should check your own jurisdiction. It is general information, not legal advice.
The inbox that never empties
A property manager with a rent roll of a couple of hundred properties does not run out of messages. A tenant reports a dripping tap, another a broken heater, a landlord asks why rent is late, a tradie wants approval for a quote, and an inspection report needs writing before Friday. Most of these are routine. A small share are urgent, and that small share is the reason nobody can safely ignore the rest.
The real skill in this job is sorting fast and acting on the right ones. That is where AI helps.
AI for property management: maintenance triage
Tenants report faults by email, portal, text and phone, often with a photo and a vague description. An AI assistant can read each report, ask the missing questions (where is the leak, is water coming through the ceiling, is the power off), classify it, and create a job with the details attached.
The classification is the part that must be right, because urgent repairs have legal meaning. NSW Government guidance lists urgent repairs as including gas leaks, dangerous electrical faults, floods, major storm or fire damage, loss of gas, electricity or water supply, burst pipes, a blocked or broken toilet, serious roof leaks, and faults that make a property unsafe or insecure (NSW Government, urgent repairs, opened November 2026). The page says tenants must tell the landlord or agent immediately. A request that sits in a queue for two days is a problem for the agency.
So the design is conservative:
- When the AI is unsure, it escalates to a person instead of marking a job as routine.
- Keywords and photos suggesting gas, electrical, flooding or security risk alert a person immediately, in or out of hours.
- It tells the tenant what happens next and when, without promising a tradie’s arrival time it cannot control.
- Approvals above the owner’s limit still go to the owner or a person with authority, with the quote and photos attached.
Whether a repair is the landlord’s or tenant’s responsibility can be disputed. The AI records the facts; a property manager decides. For the phone side of this, see AI receptionists for property management, and for the wider process, property management automation.
Arrears reminders
Most late rent is a missed payment, not a crisis. A polite reminder on day one or two, a second the following week and a message offering to talk if something has changed will solve many cases before they become formal. AI can write these in a calm tone, personalise them to the lease, and answer “how much do I owe” from the ledger.
The line is where informal becomes formal. In New South Wales, NSW Government guidance says a landlord or agent can serve a termination notice when a tenant is more than 14 days behind with the rent, giving 14 days to leave. The notice must be in writing, signed and dated, state the address, give the day the tenant must move out, state the reason, and tell the tenant they do not have to leave if they pay all the rent owing or agree to and fully follow a repayment plan (NSW Government, non-payment of rent, opened November 2026). Other states have different periods and forms.
A property manager should prepare and sign any notice after checking the ledger, because a wrong amount or date can invalidate it and a wrong notice to a tenant who has actually paid causes real harm. Use AI to flag who is approaching the threshold, never to send notices automatically.
Rules for AI-written arrears messages
- Amounts and dates come from the ledger, not from the AI’s wording.
- No threats or statements about legal consequences unless a person approved them.
- Hardship replies go to a person, with the tenant’s message attached.
- Every message gives a way to reach a human.
Inspection reports
Routine inspections generate photos and rough notes. AI can turn those into a structured report by room, caption photos, list maintenance items and draft the owner summary. It is a time saver of an hour or more per week for busy managers.
It is also where invented detail shows up. The AI may describe a wall as freshly painted that merely looks clean, or call a mark “pre-existing” without evidence. Condition reports can matter in a bond dispute, so the property manager checks each statement against the photos, and anything the AI could not see is marked as not assessed. Tell tenants how reports are produced and keep their photos secure. Our guide to AI document processing shows how documents are extracted and checked.
Owner communication
Monthly owner updates are mostly assembled from facts already in your software: rent received, jobs completed, upcoming lease dates. AI can write them in a readable form. Add the manager’s own comment where judgement is needed. Owners value a call about a difficult tenant more than a perfect template.
What stays with the property manager
| Job | AI | Property manager |
|---|---|---|
| Maintenance requests | Sorts, asks questions, creates the job | Confirms urgent items and approves spend |
| Rent reminders | Drafts and sends the early ones | Handles replies about hardship |
| Termination and breach notices | Flags candidates | Prepares, checks and signs |
| Inspection reports | Drafts from notes and photos | Verifies every statement |
| Disputes and tribunal matters | None | All of it |
Privacy and accuracy
Tenants and owners give you IDs, incomes, bank details and photos of their homes. Keep this inside your property management software or a tool under a written business agreement, and do not paste it into free chat tools. The privacy checklist is in AI and the Privacy Act. To reduce invented details, restrict the AI to the ledger, lease and notes it is given; our guide to reducing AI hallucinations sets out how.
A worked example: a Saturday-night leak
A tenant texts at 9pm on Saturday: water is coming through the bathroom ceiling. An assistant asks whether the water is near any light fittings or power points, whether anyone can reach the stopcock, and for a photo. The words “ceiling” and “water” plus a possible electrical risk place it in the urgent category, so it notifies the on-call manager immediately and tells the tenant to turn off power at the board only if it is safe to do so, as per your standard instructions. A routine request such as a sticking door goes into Monday’s queue with a note to the tenant. The manager decides who is called out on the night.
The value is the sorting. Nobody scrolls the inbox on a Sunday to find the one message that matters.
Lease renewals and routine notices
Lease expiries, rent review reminders and routine inspection notices follow a calendar. Software can list those due in the next 60 to 90 days, draft the communication to the owner and the tenant, and keep track of replies. The required notice periods and forms depend on the state, so build them into the templates from your regulator’s current forms and have a manager sign off, rather than letting the AI choose the wording.
Where to start
- Put a weekly count on how many maintenance and arrears messages arrive and how long each takes to answer.
- Start with triage and acknowledgement, with escalation rules written by your most experienced manager.
- Add early arrears reminders for a small group of properties.
- Try inspection report drafting on one manager’s portfolio, with a checking step.
CentriWeb sets up automation that connects to the property software you already use, for property businesses. If your software has built-in reminders and request forms you are not using, that is a cheaper first step.