Tensor Planet
Customer case study · Waste · Portland, Maine

How Troiano Waste Services got ahead of exhaust failures.

Preventive maintenance was already in place. The telematics were already running. Exhaust and aftertreatment failures still arrived as derates in the middle of a collection route, and the same units kept coming back through the shop. This is what changed when the data already coming off those trucks started predicting the failure instead of reporting it.

Fleet
55 heavy-duty collection vehicles
Based
Portland, Maine. Hauling since 1976.
Running
Samsara telematics and CMMS repair history
In scope
Exhaust and aftertreatment systems
A Troiano Waste Services front loader parked at the yard
The outcome

What the engagement produced.

41%

fewer exhaust-related repairs, in the first 10 weeks

$18,184

recovered in those 10 weeks, across 55 trucks

$87,642

realized savings at the one-year review

Source: Troiano Waste Services, Maine

First 10 weeks

Measured across the pilot, on 55 trucks.

Exhaust-related repairs
down 41%
Direct repair costs avoided
$6,984
Lost revenue avoided on prevented disruptions
$11,200
Total recovered $18,184
Twelve months

Measured at the engagement review.

Per truck, per year, across 55 trucks
about $1,600
Service disruptions avoided
4+ a week
Basis
Their own repair history
Realized savings $87,642
01

The challenge

Troiano was not short of data. Soot built up faster than passive regeneration could clear it, so DPF clogging kept recurring on the same units. The failures landed as derates on a collection day, which meant a roadside call, a missed route and rework in the bay.

Clogged filters also put EPA compliance at risk, and every unplanned repair drew on a technician pool the whole industry is short of. The alerts the shop already had told them a fault had happened. Nothing told them which truck was about to have one.

Before Tensor Planet
  • 150+ exhaust incidents a year
  • 10%+ of annual maintenance spend
  • ~30 exhaust issues in a peak month

Source: Troiano Waste Services, Maine. The fleet position before the engagement.

02

The approach

Read-only on the data Troiano already had. Nothing went on the truck, and nothing changed about how the shop runs its day.

01

Data integration

We ingested the telematics feed, the repair history and the contextual fleet data directly, read-only. No manual prep, no new hardware, no export work for their team.

02

Physics-based AI modeling

The models track filter regeneration and soot buildup by vehicle type and duty cycle, alongside trouble codes. An alert fires only when the physics and the codes agree.

03

Alerts the shop can act on

Daily risk alerts arrive aligned to the PM schedule, with the recommended action for each vehicle. Technicians bundle the corrective work into planned shop time.

A Troiano Waste Services technician crossing the shop floor between service bays
Repairs moved from the roadside into the bay.
In their words

Hear it straight from the fleet.

For the shop, the biggest win was how simple this was for the technicians. They didn’t need to learn a new tool or change their routine.

Scott Lane Fleet Manager, Troiano Waste Services

Tensor Planet’s system connected directly with the data we already had, without requiring any manual prep work or new hardware.

Abel Cota Director of Information Systems, Troiano Waste Services

This initiative fits perfectly with our focus on innovation and efficiency. The financial results speak for themselves.

TJ Troiano COO, Troiano Waste Services

Watch on YouTube

What it is like knowing about a breakdown before it happens.

Abel Cota, Director of Information Systems, Troiano Waste Services

Watch on YouTube

How complex the integration really was.

Abel Cota, Director of Information Systems, Troiano Waste Services

As featured in
  • MIT Sloan Management Review
  • Forbes
  • Transport Topics
  • Waste Dive
  • Waste Advantage

See which of your vehicles will break down next.