Tenant operations AI for a 200-unit Dubai portfolio with a 90-second response time

Industry

Property Management

Market

UAE

Client

Residential property manager in Dubai

Sponsor

Managing Director

What we built

Tenant operations AI

The situation

The client manages a 200-unit residential portfolio in Dubai. For the Managing Director, the problem was volume, not complexity: every maintenance call, every lease renewal and every tenant question landed on the same small team. Leaking taps, access cards, parking, cheque schedules, renewal terms, move-out inspections. Individually simple, collectively relentless.

The team spent its days answering and chasing rather than managing: returning calls, finding the right contractor, reminding tenants about renewals, and repeating the same answers to the same questions. Tenants waited, and the portfolio's owners got a team with no time to improve anything.

What we built

An AI that handles tenant communication end to end and turns requests into tracked work.

One front door for tenants. Tenants reach the AI by phone or message, in English or Arabic, at any hour. It recognises the tenant and the unit, so it can answer about their lease rather than leases in general.

Maintenance triage and dispatch. The AI asks the questions a property manager would (what is broken, where, since when, is water or power involved, is it an emergency), classifies urgency, and creates the work order. Emergencies such as leaks, power loss or lockouts are escalated immediately to the on-call contact. Routine jobs go to the right contractor category with the details and the tenant's availability attached. The tenant gets updates as the job moves.

Lease renewals that start on time. The AI tracks renewal dates, contacts tenants ahead of expiry, explains the renewal terms the manager has approved, captures intent to renew or vacate, and hands the documentation step to the team with everything prepared. Anything involving negotiation goes to a person.

Answers from the portfolio's own information. Building rules, amenity hours, payment schedules, move-in and move-out procedures, and community requirements are answered from the manager's own documents, not generalities.

A clean record. Every conversation, request and outcome is logged against the unit in the property management system, so the team sees history at a glance.

How it went live

The AI started with the highest-volume requests (maintenance and routine questions) and took on renewals once the team was confident in the triage. Emergency handling was tested explicitly before launch: what the AI says, who it calls, and how quickly.

Results

Tenants now get a response in 90 seconds, whatever the hour. The AI handles maintenance calls, renewals and tenant queries across all 200 units. The Managing Director's description of the change for the team was that they went from reactive to strategic overnight: less time answering the phone, more time on owners, vacancies and the portfolio itself.

What made the difference

Requests become work orders. The value is not in answering the tenant; it is in the job being created, routed and tracked without anyone retyping it.

Emergencies designed first. Getting the urgent path right is what let the manager trust the AI with everything else.

Knowing the tenant. Recognising the unit and the lease turns a generic assistant into a property manager's assistant.