Property managers should let AI read each tenant message, in any channel or language, and turn it into a work order with the unit, the issue, photos and urgency. It routes the order to the right technician or contractor, drafts status updates for the tenant, and leaves emergencies, disputes and anything that costs the owner to a named person.
#Key takeaways
- Tenants complain less about the repair than about not knowing what is happening with it.
- The work starts with intake: one record per request, whatever channel it came in on.
- AI can classify, route and draft updates; people decide on emergencies, owner approvals and disputes.
- The same records feed the owner report, so nobody rebuilds it from WhatsApp threads.
- A property management system with a tenant portal may already cover a single building; a workflow pays off across a portfolio with mixed channels.
#What does a property manager's inbox look like?
A tenant sends a WhatsApp voice note in Arabic about a leaking AC. Another emails the building manager, a third calls the security desk, and a fourth posts in the building group. The property manager copies each one into a spreadsheet or the maintenance system, calls the contractor, and then answers the same tenant who asks for an update twice.
At the end of the month the owner wants a report: what broke, what was fixed, what is open, what needs their approval. The property manager rebuilds it by scrolling chats.
#Why do slow updates hurt more than slow repairs?
Tenant surveys keep putting communication at the top. An AppFolio survey reported by Multifamily Dive, reporting an AppFolio survey, checked november 2026 found lack of communication was the top tenant complaint, ahead of maintenance delays. The Buildium 2026 renters report, checked november 2026 ranks fast, reliable maintenance responses first and clear, consistent communication second.
Most tenants accept that a part takes time to arrive. What they do not accept is silence. That makes status updates the cheapest place to improve, and the place where AI drafting helps most.
#A maintenance intake playbook
- Pick one intake record. Every request, from any channel, becomes a work order in your maintenance or property management system, never a chat thread only.
- Write your categories and urgency rules: what counts as an emergency (water, power, lifts, security), who gets each category, and what needs owner approval first.
- Let AI read the message, photos and voice notes, fill the work order and suggest category and urgency. The coordinator confirms until the suggestions stop needing edits.
- Route emergencies straight to a person by phone, not a queue.
- Draft tenant updates at each status change, in the tenant's language; the coordinator approves or edits.
- Close the loop: ask the tenant whether the issue is fixed before the order is closed.
Measure first acknowledgement time, open orders by age, repeat requests for the same issue, and the share of tenants who had to chase. Measure before you start.
#How does the owner report come out of the same records?
Once every request is a work order with a category, status and cost approval, the monthly owner report is a summary of records, not a reconstruction. AI drafts the narrative per property: what was fixed, what is open and why, what needs the owner's decision. The property manager checks it and sends it.
This is often the second workflow, not the first. It is only as good as the intake records behind it.
#When is a tenant portal enough?
If you run one or two buildings, tenants already use the portal of your property management system, and requests rarely arrive elsewhere, turn on the portal's request forms and automatic notifications first. That may be all you need.
A workflow earns its place across a portfolio where requests arrive by WhatsApp, email and phone in Arabic and English, contractors vary by building, and owners expect reports you currently assemble by hand. See also AI for ticket triage.
#Where does Stennir fit?
Stennir starts with AI consulting: the AI Opportunity Audit ranks five workflows, here typically maintenance intake, owner reports, renewals, collections follow-up and leasing enquiries, each with a cost and payback case, and recommends one build. AI implementation builds it into your maintenance system and WhatsApp, with the coordinator approving.
AI training then works with coordinators and building managers on the real queue. The gap is familiar: JLL, october 2025 found most real estate firms piloting AI, but only 5% had achieved all their AI goals.
We are
Stennir maps your maintenance intake, builds routing and tenant updates into the systems you already use, with a named person approving, and trains the team that runs it.
We aren't
Stennir is not a property management system, a facilities contractor, or a bot that tells tenants a repair is done when it is not.
#What should a head of property management do next?
Take last month's requests and count how many arrived outside your maintenance system. That is your baseline. See AI for operations teams, then take the free readiness assessment or book a 30-minute call.
#Common questions
Can it understand voice notes and Gulf Arabic?
Yes. Arabic and English are fully supported, including Gulf dialect, mixed-language messages and voice notes, which are transcribed before the request is classified. Stennir tests the intake on a sample of your own past requests first, and anything the system is unsure about goes to the coordinator instead of being guessed.
Does AI approve repairs or spend?
No. AI suggests a category, urgency and contractor, and drafts updates. Approvals that commit the owner, contractor choices outside the agreed list and every emergency stay with a named person. The approval rules are written with your team before anything is connected, and every action is logged.
What stays with us?
Everything that matters: the accounts, the tenant data, the work order history, the categories and rules, and the off switch. The workflow runs in your systems and hosting region, so if you stop working with Stennir, the records and the process stay in place.
What could go wrong?
An urgent leak classified as routine, a request routed to a contractor who has left, or an update that promises a visit nobody booked. That is why emergencies go to a person, contractor lists come from your system, updates are approved before sending, and wrong classifications are reviewed each week.
