Which Nearby Car Can the Customer Actually Take?

The nearest dot on the map

Distance is the easy part. A car sharing app sorts cars by coordinates, but those coordinates are only as good as the last message the car sent. Under a mall or a tower, that message can be hours old. And a car can be close and still useless to this particular customer for half a dozen ordinary reasons.

Here’s what we’d check before telling anyone to walk over:

• When the car last reported its position. A car out on the street reports as often as your installation is set up to; one on level P3 may have been silent since lunch.
• Whether a rental is already running. Someone may have taken it two minutes before your customer opened the app.
• Who is allowed to see it. Corporate cars, cars for a hotel’s guests or a zone shown only to one client group all look like ordinary cars on an ops map.
• Open technician tasks and alarms. A car with a damage report can be hidden from booking until it’s fixed, and a car with a tracker task may not be where it seems.
• Fuel, if your telematics reports it. A customer heading to Abu Dhabi on 15% isn’t a happy customer.

Support usually checks these one screen at a time: the map, then the rental, then the tasks, then the customer’s group. With a queue of calls on a Friday evening, that’s where the minutes go.

Friday evening, one question

The support agent has ChatGPT connected to the platform and types the way people type on a busy shift:

Support agent: customer at dubai mall, fashion avenue parking. app shows a car 200m away but booking fails. which cars within 1 km can they actually take right now? closest first, and tell me why the others are out. read only

ChatGPT: #262 is the closest car this customer can take. It’s parked on the Downtown side, about a ten-minute walk. #318 is probably what the customer sees: its last position was recorded before it went quiet, so it may already have been moved. I’ve listed #331 as well: it’s just outside the 1 km you asked for, but it’s the next car they could take.

The agent asks one more thing before calling the customer back:

Support agent: who has the 318 task and has anyone started it? and does 331 have a refuel task?

ChatGPT: The task on #318 was created at 16:05 and assigned to the Downtown technician. It hasn’t been started. #331 has no open task.

Three minutes, and the agent knows what to say, which car to point to and where the real problem sits. If something in the table looks off, #318 is one click away in the control panel.

Operator console · demo data

The fleet data the AI assistant answers from

Fleet overviewLive map

Fleet overview. Utilisation, active rentals, revenue in AED and the cars that need attention: the figures the assistant reads when you ask about your fleet. Technicians get a prioritised task list.

Live map. Every vehicle, its status and service zones on one map, updated in real time. Riders see the same available cars nearby in their app, with the model and fuel level.

×

Who decides what

The assistant ranked and explained. The rest was people.

The agent told the customer to book #262, gave them the level and the walking direction, and stayed on the line until the booking went through. The customer booked it in the app themselves. Then the agent asked the assistant to create a refuel task for #331 in the Downtown zone, saw the preview, pressed Confirm, and the task went to the field team. Nobody touched #318: a task was already open, and a second one would only confuse the technician. The shift lead got a one-line note to chase it.

On control: Ask your fleet, the MCP server built into CT Mobility, works under each employee’s own login and permissions and is read-only by default. Changes work only if the employee allowed them when connecting, each one needs a confirmation, and every call is written to the audit log. We enable it per installation, on request.

Where the map goes wrong

Start with the position timestamp: if it’s much older than your installation usually reports, treat the dot on the map as a guess. Underground car parks at Dubai Mall, DXB Terminal 3 or the towers in JLT swallow the signal, and the app keeps showing the last place the car reported. Booking status and customer groups come next. Fuel matters only where the installation reads it from telematics; if it doesn’t, the assistant says so instead of guessing.

Airports are their own case. At DXB Terminal 3, a customer off a late flight wants the car closest to arrivals, and the car closest on the map can be one floor away in a car park where the signal is weak. The same question works there, with one change worth making: ask for cars whose position is under 15 minutes old first, then the rest. A walk to a car that turns out to be somewhere else costs more at an airport than anywhere in the city.

The opposite call happens too: a customer who can’t end a trip in the same basement. We wrote that one up in a customer couldn’t end a rental.

Tomorrow morning

1. Give support one question to ask for “car won’t book” calls: the cars within a kilometre the customer can take, closest first, with the reason for each one that’s out.
2. Agree what “too old” means for a position. Thirty minutes? Two hours? Write it down.
3. List the car parks where your cars go quiet, and check whether your zone map treats them differently.
4. Ask the shift lead to look at open “check tracker” tasks before the Friday evening peak.

For the platform behind this, see the AI car sharing platform. If you’re weighing AI tools for your back office, start with our guide to MCP for car sharing and rental operators.

CT Mobility Solutions F.Z.C. · E-Commerce Hub BC-8, Ajman Free Zone C1 Building, Ajman Free Zone, Ajman, United Arab Emirates

Frequently asked questions