AI use case · Property

AI maintenance triage: the right technician, the right priority.

For property and facilities teams where a coordinator reads every request, decides the trade and urgency, and phones round for a technician. AI sorts each request and suggests who should go and when; the coordinator confirms.

Ahmed Khaled, FounderLast updated:

Discuss this use case

A short note is enough. Phone and email let us reach you.

About: AI & Intelligent Automation

Tell us about the workflow, who uses it and what is difficult today. Please leave out passwords, confidential records and sensitive personal information.

Add company and other details (optional)

After you send this, choose a time for a free 30-minute discovery call. If no time suits, we get back to you as soon as possible to arrange one.

What is AI maintenance ticket triage?

AI maintenance triage reads each incoming request, from chat, app or email, and classifies it by trade, urgency and location. It suggests a technician or contractor and a visit slot based on your rules and schedules. A coordinator confirms or changes the suggestion before the work order is issued and the tenant is updated.

The problem

Every request is sorted by hand.

In UAE buildings, maintenance requests arrive through every channel at once: the tenant app, WhatsApp, email, the reception desk and phone calls. In summer, AC complaints alone can fill a coordinator's day. Each request has to be read, given a trade and priority, and matched to an in-house technician or a contractor.

When this depends on one person, urgent jobs wait behind routine ones and the same issue is logged twice. Contractor visits are booked by phone and the tenant is not told what happens next. Later, nobody can easily say why a job was late or which building generates the most calls.

  • Manual sorting

    A coordinator reads every request to decide trade, urgency and location before anything moves.

  • Mixed priorities

    A water leak and a loose cabinet door can sit in the same queue until someone notices the difference.

  • Silent tenants

    Requests are assigned by phone, so tenants chase for updates and the record is incomplete.

How it works

One maintenance request, from report to scheduled visit.

An illustrative workflow. The categories, priority rules and target system are agreed for each project.

  1. Request

    A request arrives through the tenant app, WhatsApp, email or a form, with text and photos where the tenant sends them.

  2. Classification

    AI identifies the trade, such as AC, plumbing, electrical or civil, the unit and location, and any duplicate open ticket.

  3. Priority

    Your priority rules are applied, with words like leak, no power or no cooling raising urgency. Safety cases go straight to a person.

  4. Suggested routing

    It suggests an in-house technician or contractor and a visit slot, using the building, trade, maintenance contract and schedule.

  5. Coordinator check

    The coordinator sees the request, the suggestion and the reason, then confirms or changes it.

  6. Update

    The work order is issued in your maintenance system and the tenant receives the reference and the planned visit.

Who does what

What the AI does · What a person checks.

What the AI does

  • Reads requests in Arabic and English, with photos where available

  • Classifies trade, location and likely urgency

  • Spots duplicates of an open ticket for the same unit

  • Suggests a technician or contractor and a visit slot, with the reason

  • Drafts the tenant update for the coordinator to send

What a person checks

  • Every routing suggestion, before the work order is issued

  • Any request marked urgent or with a safety risk

  • Requests that may be chargeable to the tenant or the owner

  • Changes to priority rules and contractor lists

First pilot

A first pilot on your own maintenance history.

We test on your past requests before anything goes live. Scope, timeline and price are confirmed in a written proposal.

  • One portfolio or building group

    Start where request volume is steady and the trades and contractors are known.

  • Your past tickets

    A sample of closed requests, shared safely after the first call, with the trade and priority your team actually gave them.

  • Written rules

    Your priority definitions, escalation rules and contractor coverage, agreed in writing before testing starts.

  • Side-by-side running

    The triage runs next to your coordinators first, and its suggestions are compared with their decisions before anyone relies on it.

Data, hosting and Arabic

Where the data goes, and how Arabic is handled.

  • Hosting

    The application and stored data can be hosted on Azure UAE North, on private servers managed by Vloud, or in your own cloud account. The AI model is used through a provider agreed per project, and its processing region may be outside the UAE.

  • Access

    Coordinators see requests for their buildings. Contractors see only the jobs assigned to them and the details needed to do the work.

  • Arabic

    Arabic, English and mixed requests, including short dialect messages, are tested as their own sample.

  • Photos

    Tenant photos are stored with the ticket. Retention and who can view them are agreed with you.

Honest limits

Where it does not work yet.

  • It suggests; it does not dispatch on its own. A coordinator confirms every routing decision in the first phase.
  • Safety emergencies such as fire, gas smell or lift entrapment are not triaged by AI. They follow your emergency procedure.
  • Vague requests like "something is wrong" need a follow-up question or a person before they can be classified.
  • We do not quote accuracy or time savings before the pilot. Suggestions are measured against your coordinators' own decisions.
  • Slot suggestions are only as good as the schedules they read. If technician calendars are not kept in a system, routing stays manual.

What drives cost and timeline. The number of channels, buildings and trades, the priority and escalation rules, how many contractors and maintenance contracts are involved, the maintenance system's integration route and hosting requirements. We don't publish a price list; scope, timeline and price come in a written proposal after the discovery call.

Evidence

What we can show you.

Every example is labelled: delivered work, our own product, a demonstrator, or an illustration with synthetic data.

  • Our product

    ArkanPM maintenance and work orders

    ArkanPM is our own property management product. Its maintenance and work orders module follows a request from the first report through assignment to completion. It is a product we own; this triage would be custom work on top of it or your current system.

  • Delivered work

    EVS customer apps and staff portal

    We delivered Arabic-first iOS and Android apps and a staff web portal for EVS, integrated with their workshop system. In the EVS launch period, 4-18 October 2026, 31% of app bookings were made between 6 pm and 6 am. It is service booking work, not an AI triage project.

Before you start

Questions buyers ask.

Which channels can requests come from?

A tenant app, WhatsApp, email or a web form. Each channel is connected in the pilot only if it is part of the agreed scope, and all of them feed one queue.

Does it send technicians automatically?

Not at first. It suggests a technician and a slot with the reason, and a coordinator confirms. Automatic dispatch for low-risk jobs can be considered later, if your team agrees after the pilot.

How does it decide what is urgent?

From priority rules your team writes, such as active leaks or no cooling being urgent. The AI applies those rules to each message and shows why it chose a priority.

Does it work with ArkanPM or our current system?

It can create work orders in ArkanPM, or in another system with a supported API or an agreed import route. We review your system before confirming the integration scope.

Can contractors use it?

Contractors can receive their assigned jobs with the details they need. What they see and how they update the job is agreed with you.

Where is the data hosted?

The application and stored data can be hosted on Azure UAE North, on private servers managed by Vloud, or in your own cloud account. The AI model is used through a provider agreed per project, and its processing region may be outside the UAE.

Who owns the code?

You own the source code and IP of the custom work we build for you. Our own products, including ArkanPM, stay ours, and any licence terms are agreed separately.

Next step

Test it on your own documents.

Tell us what the work looks like today. We reply with the questions we need answered before a pilot can be scoped.

Discuss this use case

A short note is enough. Phone and email let us reach you.

About: AI & Intelligent Automation

Tell us about the workflow, who uses it and what is difficult today. Please leave out passwords, confidential records and sensitive personal information.

Add company and other details (optional)

After you send this, choose a time for a free 30-minute discovery call. If no time suits, we get back to you as soon as possible to arrange one.