Fleet operator reduces missed-service call-back processing time by 38%

A regional fleet operator running a few hundred trucks across three yards was losing service requests in the gap between a missed call and a call-back. The requests were not lost for good. They were lost for hours, and hours were enough to turn a routine repair into a roadside breakdown. Here is how they cut the time it took to process a missed-service call-back by 38 percent with a Torque agent doing the first pass and the service desk handling only what needed a person.

Where it started

The service desk had two coordinators covering three yards and a phone line that rang all day. When both were on a call, on the radio, or out at a bay, inbound calls went to voicemail. Drivers, yard leads, and customers left messages about check-engine lights, brake noise, a lift gate that would not cycle, a unit that failed a pre-trip. Every one of those messages had to be listened to, understood, matched to a unit number, logged in the maintenance system, and returned with a call.

The team tracked it for a month. The average missed-service call-back took just over three hours to process from the time the voicemail landed to the time a work order existed and the caller knew what would happen next. Some took ten minutes. Some took until the next morning. The ones that took until the next morning were the ones that hurt: a truck that could have been pulled and fixed at the end of a shift was instead on the road at 5 am with the same problem.

Nothing about this was a people problem. The coordinators were good and busy. It was a queue problem, and the queue only got worked when someone had a free hand.

What changed

Torque put a service agent on the line. It does not replace the coordinators. It takes the part of the job that is listening, typing, and looking things up, and leaves them the part that is judgment.

• Every missed call is answered by the agent, not by voicemail. It asks for the unit number, what the driver is seeing, where the truck is, and whether it is safe to keep driving.
• The agent looks the unit up in the maintenance system, pulls the open defects and the last service date, and drafts a work order with the right priority code.
• Routine requests are scheduled on the spot against bay availability and the caller gets a text confirmation with a time and a bay.
• Anything with a safety flag, an unfamiliar symptom, or a unit already under an open work order is routed to a coordinator with the full transcript and the draft work order attached.

The coordinators still see everything. The difference is that they see it as a queue of prepared work orders to approve or adjust, not as a voicemail box to work through between calls.

A Tuesday, before and after

Before: a driver calls at 2:10 pm about a brake warning light. Both coordinators are busy. The voicemail is heard at 3:40 pm. The coordinator calls back and gets no answer because the driver is unloading. They connect at 4:25 pm, the unit is looked up, the work order is written at 4:50 pm, and the truck is scheduled for the next day because the bays are full by then.

After: the same call at 2:10 pm is answered by the agent. The driver gives the unit number and describes the light. The agent flags it as a brake item, checks the unit, sees no open work order, and routes it to a coordinator as a priority. The coordinator approves the draft at 2:18 pm. The driver has a text by 2:20 pm telling them to bring the unit to bay two at the end of the route. The truck is fixed that evening.

The results

• Missed-service call-back processing time down 38 percent, from just over three hours to under two, measured across every yard for the first full quarter.
• Requests that waited overnight fell by more than half, because scheduling now happens on the first call instead of the second or third.
• Coordinators spend their time on approvals, exceptions, and the phone calls that need a person. Voicemail volume on the service line dropped to near zero.
• Every request now has a transcript, a timestamp, and a work order, which made the monthly maintenance review a report instead of a reconstruction.

How it went live

The first two weeks were spent listening. Torque worked through a sample of past voicemails with the coordinators to learn how requests were described, which symptoms were treated as urgent, and how the maintenance system expected work orders to look. The agent was configured with those rules, then run in the background for a week, drafting work orders that the coordinators compared to their own.

Once the drafts matched, the agent took the line for one yard. Two weeks later it took all three. The coordinators kept the ability to pull any call back to a person at any point, and they used it often in the first month and rarely after that.

What they would tell another fleet

Start with the queue that gets worked last. For this operator it was missed calls, because a ringing phone always lost to the person standing at the desk. Measure how long that queue takes today, put an agent on the first pass, and keep your people on the decisions. The 38 percent came from the hours nobody was counting.