Logistics / Operations45 people5 months

How a Logistics Company Cut £211k in Annual Op Costs Without Losing a Single Employee

Five workflow automations that reclaimed 300+ hours per month — and compounded every quarter as we layered in more.

£211k

saved annually

£211k

annual cost saving

300+

hours recovered monthly

8 mo

payback period

62%

ops cost reduction

0

redundancies made

The problem

Their cost structure was fine at 20 customers. At 200, every new customer made the problem worse.

Freightpath had grown 4× in two years. Revenue was up. Margins were down. Each new customer added operational complexity that required proportional headcount. The team was doing high-volume, rule-based work — things that follow predictable patterns — but they were doing it manually, one task at a time.

📅

Manual dispatch scheduling consuming 3 hours daily across two coordinators

£62k/year in coordinator time on scheduling alone
🔄

Data entry between 3 disconnected systems — TMS, CRM, and billing

14 hours/week, £38k/year in data reconciliation
📄

Invoice and contract processing done entirely by hand

400+ documents/month, 8 hours/week
📊

Weekly ops report manually compiled from 5 data sources every Friday

6 hours of senior ops time weekly
The solution

Five agents. One orchestration layer. Fully integrated with existing stack.

We didn't replace their software. We wrapped intelligence around it. Every agent integrates with their existing TMS, CRM, and billing system via API. No data migration. No new UI to learn.

Layer 01

Dispatch Intelligence Agent

Reads incoming load requests, cross-references driver availability, route constraints, and vehicle capacity. Generates optimal dispatch schedule and sends assignments. Coordinator reviews exceptions only.

Claude 3.5TMS APICustom dispatch logic
Layer 02

Data Sync Orchestrator

Real-time sync across TMS, CRM, and billing. When a job is created in TMS, customer record updates in CRM, invoice draft creates in billing — simultaneously. Zero manual data entry.

Make.comREST APIsWebhook mesh
Layer 03

Document Processing AI

Extracts data from inbound invoices, contracts, and delivery confirmations. Categorises, validates against expected values, and routes. Exception rate: 4%. Human reviews exceptions only.

Document AIGPT-4o visionAirtable
Layer 04

Autonomous Reporting Engine

Weekly ops report compiled from all 5 data sources and delivered at 7am Monday. Includes anomaly flags with plain-English explanation. Friday 6-hour report: eliminated.

GPT-4oData connectorsSlack/Email
Layer 05

Email Triage & Routing Agent

Classifies all inbound operations email by type, urgency, and required action. Routes to correct owner with context summary. Drafts responses for routine request types.

Claude 3.5Gmail APIHubSpot
Execution timeline

12 weeks. Phased delivery. Live operations throughout.

01
Week 1–2

Workflow Audit & Prioritisation

  • All 5 workflows documented in full
  • Cost-per-task baseline established
  • API access confirmed across all systems
  • Automation priority matrix built
Prioritised build order approved
02
Week 3–5

Dispatch + Data Sync Live

  • Dispatch agent deployed and tested
  • TMS-CRM-Billing sync live
  • 2-week parallel run with manual process
  • Exception threshold tuned
£80k annualised saving confirmed
03
Week 6–9

Document AI + Reporting Engine

  • Document processing AI trained
  • 4% exception rate achieved
  • Autonomous reporting live
  • Friday report process retired
£130k annualised saving confirmed
04
Week 10–12

Email Agent + System Handover

  • Email triage agent live
  • Full system monitoring dashboard
  • Team trained on exception handling
  • 30/60/90 expansion roadmap delivered
£211k annualised saving confirmed — handover complete
The results

Costs cut. Team intact. Margin expanded.

All financial figures audited against pre-engagement payroll and operational cost records. Hours data from task tracking system. Exception rates from system logs.

DimensionBeforeAfterChange
Annual ops cost£340,000£129,000
–62%
Hours recovered monthly0300+
+300 hrs
Dispatch scheduling time3 hrs/day15 min/day (exceptions)
–92%
Manual data entry14 hrs/week0 hrs/week
–100%
Document processing accuracy96% (human)99.2% (AI)
+3.2pp
Headcount45 people45 people
No change
Gross margin31%44%
+13pp
TR

We were in a trap — every new customer cost us more in ops than they should have. Aroluxa broke that relationship. Now we can onboard 50 new customers and our ops cost barely moves. That's a different business model. That's what they built us.

Tom RichterCOO, Freightpath
Strategic learnings

The principles behind what worked.

01

Automate the highest-cost constraint first

We could have started with the email agent — it was the fastest to build. We started with dispatch because it was the highest single cost. The sequencing of automation matters as much as the automation itself.

02

Wrap intelligence around existing tools — don't replace them

Every enterprise automation failure we've studied involves replacing existing systems. Teams resist it. Data migrates badly. We built on top of what was there. Adoption was immediate because nothing changed except the work disappeared.

03

The compounding effect is real — plan for it

At 12 months, we added three more automations. Each additional workflow costs 70% less to build because the infrastructure is already in place. The companies that plan their automation roadmap to month 24 — not month 3 — see the most dramatic margin expansion.

Ready to replicate Freightpath's results?

Get the same AI system built for your business

In 30 minutes we'll map your biggest operational bottleneck and show you exactly how AI can eliminate it — no fluff, no pitch deck.

Get Your AI AuditSee Automation Plan
No retainers
Results in 90 days
Fixed-scope delivery