AI in Strata and Facilities Management: What It Actually Does in 2026
A plain-English guide to what AI really does in Australian strata and facilities management in 2026 — concrete use cases, limits, and where to start.
Artificial intelligence has gone from buzzword to working tool in Australian strata and facilities management. But there is still a lot of fog around what it genuinely does day to day — and what it can't (or shouldn't) do. This guide cuts through the hype with concrete 2026 use cases, honest limits, and practical advice for managers, owners corporations and committees who want to use AI without getting burned.
What does AI actually do in strata and facilities management in 2026?
In 2026, AI in strata and facilities management mostly does the repetitive reading, writing, sorting and summarising work that used to eat a manager's day — under human supervision. It is an assistant, not an autopilot.
The practical jobs it does well right now:
- Triages and routes maintenance requests — reads an incoming issue, works out urgency and category, and suggests the right contractor or workflow.
- Drafts communications — owner notices, meeting invitations, breach reminders, contractor emails and resident updates, ready for a human to review.
- Answers resident and tenant questions — a concierge-style agent handles common FAQs (bin nights, visitor parking, how to book the lift) and logs issues that need a person.
- Summarises long documents — by-laws, AGM minutes, engineering reports, insurance policies and contracts, distilled to the parts that matter.
- Flags compliance gaps — surfaces expired certificates, overdue inspections and missing records before they become a problem.
- Surfaces anomalies — spots the levy that doubled, the contractor invoice that looks off, or the maintenance request volume spiking in one building.
What it does not do in 2026: make binding decisions, spend money, sign off on safety matters, or replace the judgement of a qualified manager. Every meaningful action still goes through a person.
How does AI triage and route maintenance requests?
AI triage reads a maintenance request the moment it arrives, classifies it, estimates urgency, and proposes who should handle it — so nothing sits in an inbox for three days.
A typical flow looks like this:
- A resident reports "water coming through the ceiling in the car park" via app, email or phone.
- The AI reads the free-text description (and any photos), recognises it as a likely plumbing or waterproofing issue, and tags it urgent.
- It checks the building's records — is this a recurring leak? Is there a preferred plumber with current insurance?
- It drafts a work order and suggests the contractor, the priority and a first response to the resident.
- A human reviews and approves before the contractor is dispatched.
In Orveya, this triage is handled by an assistant called Theo, which proposes the category, urgency and routing but never dispatches a contractor or commits funds on its own — a manager confirms first. The value is speed and consistency: every request gets read, categorised and prioritised the same way, even at 11pm.
The honest limit: AI can misjudge urgency if a description is vague ("the thing in the lobby is broken"). That is exactly why a human stays in the loop — to catch the edge cases the model gets wrong.
Can AI draft owner and resident communications?
Yes. Drafting communications is one of the most reliable, time-saving uses of AI in strata today — provided a person reviews every draft before it is sent.
Good candidates for AI drafting:
- AGM and committee meeting notices and agendas
- Levy reminders and arrears follow-ups (tone matched to the situation)
- By-law breach notices framed clearly and neutrally
- Maintenance and outage updates ("the lift will be out Tuesday 9am–1pm")
- Contractor instructions and scope summaries
- Welcome packs for new owners and tenants
The AI handles the blank-page problem and keeps the tone consistent. The manager keeps editorial control — adjusting facts, tone, and confirming names, dates and figures. Treat AI drafts as a strong first version, never a send-ready final. Anything with legal or financial weight (a formal breach notice, a special levy explanation) deserves an especially careful human read.
What is a tenant or resident concierge agent and is it any good?
A resident concierge agent is an AI chat assistant — usually inside a building's resident app — that answers everyday questions instantly and logs anything it can't resolve. These are genuinely useful for deflecting routine queries.
What a concierge handles well:
- "When is hard rubbish collected?"
- "How do I book the rooftop or the loading dock lift?"
- "What are the visitor parking rules?"
- "Who do I contact about a noisy neighbour?"
- "My fob has stopped working" — captured as a logged request, routed to the manager.
The payoff is fewer repetitive calls and emails to the building manager, and faster answers for residents. Good systems deliver the concierge inside a branded resident app and escalate to a human whenever a question goes beyond the building's known information.
The limit worth stating plainly: a concierge should answer from your building's actual information — its by-laws, schedules and records — not from general internet knowledge. If it confidently invents rules that don't exist for your scheme, that's a failure mode. Good systems constrain the agent to verified building data and hand off to a person when unsure.
How does AI summarise minutes, by-laws and reports?
AI summarisation reads long documents and produces a short, accurate digest — turning a 60-page engineering report or two hours of AGM minutes into a one-page brief in seconds.
Where this saves the most time:
- Meeting minutes — key decisions, action items and who's responsible.
- By-laws and scheme rules — "what do our by-laws say about pets / short-term letting / renovations?"
- Engineering and defect reports — the findings and recommended actions, without the jargon.
- Insurance policies and quotes — what's covered, what's excluded, what changed at renewal.
- Contracts and tenders — obligations, term, termination rights and pricing.
A practical tip: use summaries to find what to read, not as a substitute for reading the decisive bits. For a binding by-law interpretation or an insurance exclusion that affects a claim, open the source document and confirm. AI summaries can quietly drop a qualifying clause — and in strata the qualifying clause is often the whole point.
Can AI detect compliance gaps and anomalies?
Yes — and this is one of the highest-value, lowest-risk uses. AI is excellent at watching structured records continuously and flagging what's missing, expiring or out of pattern, which is tedious and error-prone for humans to track manually.
Compliance gaps it can surface:
- Fire safety statements, anti-vibration and essential-services certificates approaching expiry
- Overdue inspections (lifts, pumps, backflow, gutters)
- Missing insurance valuations or lapsed cover
- Records that should exist for a building of a given type but don't
Anomalies it can flag:
- An invoice well outside the normal range for that contractor or category
- A sudden spike in maintenance requests for one building or asset
- A levy or budget line that jumped unexpectedly between periods
- Duplicate or near-duplicate payments
Obligations vary by state and territory — fire safety statement requirements differ between NSW, Victoria and Queensland — so an AI flag is a prompt to check, not a compliance ruling. It tells you where to look; a competent manager decides what it means. Used this way, it's a tireless second set of eyes that never forgets a renewal date.
What can't AI do — and where do the risks sit?
Being clear-eyed about limits is what separates safe adopters from the ones who get caught out.
Accuracy. AI can be confidently wrong — misclassifying a request, misreading a clause, or summarising away an important caveat. Treat outputs as drafts and flags to verify, not facts.
Judgement and accountability. AI doesn't carry duties of care, fiduciary obligations or professional accountability — a licensed strata manager does. The decision, and the responsibility for it, must stay with a person.
Data privacy and residency. Strata data is personal information under the Australian Privacy Act, covering owners, tenants, contact details and financials. Ask any vendor: Where is the data hosted? Is it kept onshore in Australia? Is your data used to train shared models? Orveya, for instance, is built and hosted in Australia — these are the questions to put to anyone you evaluate.
Over-automation. The biggest real-world risk isn't a robot uprising — it's quietly letting AI send notices or close requests without anyone checking. Keep a human approval step on anything that touches money, safety, legal standing or a resident's rights. That is what "human-in-the-loop" means: the AI proposes, a person disposes.
Where should a strata or facilities team start with AI?
Start with one low-risk, high-frequency task, prove the value, then expand. Don't try to automate everything at once.
A sensible 90-day path:
- Pick one painful, repetitive job — usually maintenance triage or drafting routine notices.
- Keep humans in the loop — every AI output is reviewed before it goes out, non-negotiable from day one.
- Check the data questions — Australian hosting, privacy compliance, and whether your data trains shared models.
- Measure something concrete — average response time on requests, or hours saved drafting comms each week.
- Add a resident concierge once the back-office basics work, to cut routine query volume.
- Layer in compliance and anomaly monitoring as a safety net once you trust the workflow.
The principle holds whatever tool you choose: let AI do the reading, sorting and drafting, and keep people firmly in charge of the deciding.
Frequently asked questions
Will AI replace strata or building managers?
No. AI handles repetitive reading, writing and sorting, but it doesn't carry the legal duties, judgement or accountability that licensed managers do. It removes busywork so managers can focus on decisions, relationships and the genuinely hard problems.
Is it safe to put strata data into an AI system?
It can be, if the vendor handles it properly. Strata data is personal information under the Australian Privacy Act, so ask where data is hosted (ideally onshore in Australia), whether it's used to train shared models, and what access controls exist. Avoid pasting owner or resident details into consumer chatbots that offer no such guarantees.
What does "human-in-the-loop" actually mean?
It means the AI proposes and a person disposes. The AI can triage a request, draft a notice or flag a compliance gap, but a human reviews and approves before anything is sent, dispatched or actioned. A well-designed assistant never commits funds or dispatches contractors on its own.
How accurate is AI for reading by-laws and contracts?
Good for getting the gist quickly, not reliable enough to be the final word. AI summaries can omit a qualifying clause or exception that changes the meaning entirely. Use them to locate the relevant sections fast, then read the source text before relying on any interpretation that has legal or financial consequences.
What's the cheapest way to try AI in a small or self-managed building?
Start with a single workflow rather than a big platform rollout — for example, AI-drafted notices or maintenance triage. Modern AU strata platforms are priced per lot (Orveya starts from $1.35/lot/month including GST with every user seat free), which makes self-managed and smaller schemes a realistic place to begin without a large upfront commitment.