Book a Demo
All Case Studies Logistics

AI Route Optimisation Cuts Fuel Costs 22% Across 1,200-Vehicle Fleet

Our research explores how a dynamic routing engine could replace static daily routes with real-time AI-optimised schedules — projecting $3.1M in annual savings and lifting on-time delivery from 77% to 95% for a large-scale last-mile logistics operator.

Scenario: Large-scale last-mile logistics operator  ·  Industry: Logistics & Distribution  ·  Est. Timeline: 12 weeks

22% Fuel Cost Reduction
18% Faster Delivery
$3.1M Annual Savings
95% On-Time Rate
01  ·  The Challenge

Static Routes in a Dynamic World

Large logistics operators typically build routes each morning based on that day's delivery manifest. Once vehicles leave the depot, there is no mechanism to adapt as conditions change throughout the day — a structural weakness with major cost and service implications.

Logistics

Static Routes Built for Yesterday

1,200 vehicles followed routes generated manually each morning with no intra-day re-optimisation. New orders, traffic incidents, customer cancellations, and failed deliveries accumulated throughout the day with no systematic response — just driver improvisation.

Fuel Costs Spiralling

Rising fuel costs combined with routing inefficiencies had pushed annual fuel spend to $14.2M. Driver feedback repeatedly identified high-traffic corridors that were avoidable with alternative routes, but the static routing process had no mechanism to respond.

Deteriorating Customer Experience

Estimated delivery windows were 4-hour blocks, and actual on-time performance had fallen to 77%. 23% of inbound call centre volume was related to delivery status queries or missed windows — a growing reputational and operational cost.

02  ·  Our Approach

A VRP Solver That Replans in Under 3 Seconds

The proposed solution is a constraint-aware Vehicle Routing Problem solver combined with real-time traffic integration — giving dispatchers dynamic, always-optimal routes while preserving full human override capability.

AI Optimisation

01 · Historical Traffic Pattern Analysis

Analysed 18 months of GPS telemetry, delivery logs, and traffic API data to build a statistical model of corridor congestion by time-of-day and day-of-week for all FastShip delivery zones. This baseline informed the initial routing cost matrix.

GPS TelemetryTraffic ModellingData Engineering

02 · Constraint-Aware VRP Solver

A Vehicle Routing Problem solver using Google OR-Tools incorporates vehicle weight/volume capacity, customer time windows, driver hours-of-service constraints, and priority tiers for same-day and premium delivery SLAs.

Google OR-ToolsVRP FormulationConstraint Optimisation

03 · Real-Time Traffic Replanning

Integrated with HERE Maps Traffic API to detect live incident and congestion events. Each replanning cycle completes in under 3 seconds for zones of up to 300 active stops — fast enough for immediate dispatcher action when a route blockage occurs.

HERE Maps APIApache Kafka3-Second Replan

04 · Dispatcher Interface & System Integration

A React dispatcher dashboard with live vehicle tracking, route performance metrics, and drag-and-drop manual overrides for exceptional cases connects to the operator's WMS for order data feeds and CRM for customer notification triggers.

ReactWMS IntegrationFastAPI

Projected Outcomes

Estimated impact based on logistics optimisation benchmarks, VRP modeling, and comparable dynamic routing deployments in last-mile delivery.

$3.1M Annual Cost Savings

Projected reduction in total fuel expenditure by ~22%, with distance per delivery falling ~18% through route consolidation and more efficient stop sequencing across the fleet.

95% On-Time Delivery Rate

Projected improvement in on-time delivery performance from a typical 77% baseline to 95% — reducing complaint call volumes and driving an increase in repeat order rates.

18% Faster Average Delivery

Projected 18% reduction in average time from depot dispatch to delivery, driven by fewer backtracking movements, better traffic avoidance, and tighter stop sequencing across all zones.

Technology Stack
Google OR-Tools Graph Neural Networks HERE Maps Traffic API Python FastAPI PostgreSQL React Apache Kafka
Start Your Project

Ready to build your own AI success story?

Every engagement starts with a thorough understanding of your data, workflows, and objectives. Let's talk about what's possible for your organisation.

  • Free discovery call, no commitment required
  • Custom AI opportunity assessment
  • Clear ROI model before any work begins
Founding Client Offer

Free AI Readiness Assessment

  • 2-hour strategy session with our AI team
  • Full audit of your data & infrastructure
  • ROI model for top 3 AI use cases
  • Written roadmap — yours to keep
Book Your Free Assessment → Contact Us