Logistics

Fleet Route Optimization Platform

A real-time routing platform saving an average of 14% on fuel costs fleet-wide.

Client

Northwind Logistics

Completed

July 21, 2025

Status

Ongoing Retainer

Services

Custom Software Development, AI Automation

Business Challenge

Northwind Logistics dispatched routes manually each morning based on driver familiarity and rough zone assignments — workable at their old size, but increasingly costly as their fleet grew past 200 vehicles and fuel prices climbed.

Objectives

  • Cut fuel spend through better route sequencing, not fewer deliveries
  • Give dispatchers a tool that adjusts routes in real time as traffic and orders change
  • Keep drivers' existing hardware — no new in-cab devices

Approach

Rather than a black-box optimizer, we built the platform around a "propose, don't dictate" model: the system generates an optimized route, but dispatchers can override any leg with one click, and the system logs why for future tuning.

Deliverables

A dispatcher web platform with live route optimization, a driver-facing mobile view (browser-based, no app install), and a savings-tracking dashboard for fleet management.

Solution

Planning

We scoped a single-region pilot (40 vehicles) before committing to fleet-wide rollout, specifically to validate fuel savings claims against real driving conditions rather than simulation alone.

Design

Dispatchers needed to trust the system before they'd stop manually overriding it, so the UI surfaces the reasoning behind each route — distance, traffic, and delivery window constraints — rather than a single optimized line with no explanation.

Development

Built on Next.js with PostGIS for geospatial routing logic and the Google Maps Platform for live traffic data, deployed as a browser-based tool compatible with the tablets already mounted in each truck.

Testing

Ran parallel routing for three weeks — dispatchers planned manually as usual while the system generated its own routes silently in the background — to compare fuel and time estimates before going live.

Deployment

Expanded from the 40-vehicle pilot to the full 230+ vehicle fleet over eight weeks, region by region, with fuel savings validated against each region's baseline before moving to the next.

Optimization

We remain on a monthly retainer refining the routing model as Northwind adds new depots and delivery zones — this is one of the few engagements still actively evolving post-launch.

The single highest-leverage change wasn't the optimization algorithm — it was letting dispatchers override routes with one click. Adoption jumped once they stopped feeling like the system was taking control away from them.

Results

14%

Fuel cost savings

230+

Vehicles routed daily

96%

On-time delivery rate

Technology Stack

Next.jsPostgreSQLPostGISGoogle Maps Platform

Gallery

Want a system like this for your business?

Tell us about your workflow — we'll scope it honestly.