Maritime · Legacy Modernization · 2020

Maritime OS

Replaced paper manifests and terminal-specific software with a unified logistics intelligence platform operations teams actually use.

Scroll to explore

← Drag or scroll →
01

Problem

1892 Infrastructure

Global cargo tracking on paper, incompatible systems, human coordinators.

130+Years operating
02

System

Logistics Intelligence

Unified platform for tracking, coordination, and operational visibility.

16Weeks to MVP
03

Process

Ops-First Design

Built with terminal operators, not for them. Adoption as success metric.

3Port integrations
04

Outcome

Legacy Modernized

Digital infrastructure matching operational reality.

↑Coordination efficiency
Client
Crowley Maritime
Engagement
Strategy + Design Lead
Duration
16 Weeks
Primary Outcome
Legacy Modernized
Stack / Tags
Ops · Enterprise · B2B

Artifacts from this engagement

ProblemChallenge BriefConstraints and context before the engagement ProcessEngagement PhasesHow the work was structured and delivered SystemSolution ArchitectureWhat was built and how it works OutcomesResults & MetricsMeasurable impact and learnings
Logistics / Product Strategy

Maritime
OS.

Crowley Maritime has operated since 1892. For most of that time, global cargo tracking ran on paper manifests, terminal-specific software, and a network of human coordinators managing information across incompatible systems. This engagement replaced that infrastructure — not with a dashboard, but with a unified logistics intelligence platform that operations teams actually use.

0
Ports Connected in Real-Time
130+
Years of Company History
47%
Ops Efficiency Gain
2022
Year Completed
Client
Crowley Maritime
Role
Product Strategy + Design Lead
Timeline
16 Weeks
Outcome
Legacy Modernized
03
Section 01
The
Problem.
A century of operational intelligence trapped in systems that could not communicate with each other.
130 Years of
Institutional Memory.
Zero of it Accessible.

Crowley's operations infrastructure was a sedimentary record of every software generation since the 1980s. Cargo tracking lived in one system. Port scheduling in another. Customs documentation in a third. Customer-facing status updates were manually compiled from these sources by a coordinator team whose primary job was information relay — translating between systems that had never been designed to communicate.

The problem wasn't that the data didn't exist. The problem was that it existed everywhere simultaneously and nowhere in particular. A cargo manager asking "where is container CSLU 4421009 right now" could expect an answer in 4 hours if they were lucky, 24 hours if the coordinator was busy, and a shrug if the container was mid-ocean on a vessel with intermittent connectivity.

01
Seven disconnected operational systems — cargo tracking, port scheduling, customs documentation, vessel management, customer service, billing, and compliance each ran on separate platforms with no real-time data exchange. Integration was manual and human-dependent.
02
Cargo visibility latency of 4–24 hours — customers and internal teams had no live cargo status. In time-sensitive supply chains, a 24-hour window between event and information is a structural failure, not an inconvenience.
03
Exception handling was entirely manual — when cargo was delayed, mis-routed, or held at customs, the process of notifying all affected parties and coordinating resolution depended entirely on a single coordinator's knowledge and initiative. No systematic escalation, no automated alerting.
04
No predictive capability whatsoever — operations ran entirely in reactive mode. Delays were discovered after they happened, not anticipated before. The cost of this reactivity compounded across thousands of shipments per quarter.
04
Section 02
The
Process.
16 weeks. Operational immersion before a single wireframe.

Understand the
Operational Reality.

Legacy modernisation fails when designers treat the legacy system as the problem. The legacy system is the solution to a real problem — the question is which parts of that solution are worth preserving and which parts can be improved by connecting them.

Key Insight · Week 3 · Port of Jacksonville
The coordinator team was not a workaround. They were the system. The new product had to make them more capable — not make them redundant. Every coordinator we replaced with automation, we lost institutional knowledge we could not recover.
01
Operational Immersion
Weeks 1–4

Embedded with operations teams at three ports — Jacksonville, San Juan, and Long Beach. Mapped every information flow, every system boundary, every human workaround. Documented 89 distinct information handoff points across the operation. The coordinator team explained what they actually did — and why the system forced them to do it that way.

02
Data Architecture Design
Weeks 5–8

Mapped the data model that could unify the seven legacy systems — not by replacing them but by building an integration layer above them. Defined the canonical cargo event model: every state a shipment could be in, every party that needed to know about it, every trigger that should produce an automatic notification or alert.

03
Interface Design + Validation
Weeks 9–13

Designed the Maritime OS interface around three user archetypes: the cargo manager (real-time fleet and cargo visibility), the coordinator (exception management and escalation), and the customer (self-service status and documentation). Tested each interface with actual users — port workers, not product managers — before finalising any screen.

04
Predictive Layer + Handover
Weeks 14–16

Scoped and prototyped the predictive exception layer — using historical delay patterns by port, route, and season to surface early warnings before events became incidents. Full design documentation, coordinator training programme, and a phased adoption roadmap that protected the institutional knowledge embedded in the existing team.

05
Section 03
The
Solution.
Four decisions that unified seven systems without breaking what 130 years had built.
01
01
Unified Cargo Event Model

A canonical data model that defined every state a shipment could be in, from booking to delivery. The seven legacy systems remained — but all events now flowed through a single integration layer, visible in real time across the organisation and to customers.

02
02
Role-Specific Interfaces

Three interfaces built from the same data model, optimised for three different jobs. The cargo manager sees a fleet-level view. The coordinator sees an exception queue with escalation tools. The customer sees their shipment, their documents, their timeline — and nothing else.

03
03
Automated Exception Alerting

The coordinator team's manual notification work was systematised — not replaced. Delay events, customs holds, and vessel schedule changes now trigger automatic stakeholder notification. Coordinators shifted from information relay to exception resolution: the high-value work the system couldn't do.

04
04
Predictive Exception Layer

Historical delay patterns by port, route, season, and carrier surfaced as early warnings in the coordinator interface. Shipments at statistically elevated delay risk were flagged 48–72 hours before the likely event — time enough to act rather than react.

Maritime OS · Real-Time Cargo Intelligence Dashboard
Jacksonville · FL
San Juan · PR
Long Beach · CA
Mid-Atlantic · Transit
Rotterdam · NL
Hamburg · DE
Active Shipments · Live
CSLU 4421009 · CG-2291
Jacksonville → Rotterdam
ON TIME
CSLU 8831024 · CG-2292
Long Beach → Hamburg
CUSTOMS HOLD
TEMU 9912301 · CG-2293
San Juan → Jacksonville
ON TIME
HLXU 2200881 · CG-2294
Rotterdam → Long Beach
48HR RISK FLAG
MSCU 5541100 · CG-2295
Hamburg → San Juan
ON TIME
Fig. 01 — Maritime OS real-time cargo intelligence view. Six active shipments, two requiring coordinator attention. Live status, 0-hour latency.
06
Section 04
The
Outcome.
What changed — in visibility, in response time, and in operational cost.
47
0%
Ops Efficiency Gain
Operational throughput improvement across all ports
24
24hr→0
Cargo Visibility Latency
From 4–24 hour delays to real-time for all shipments
7
0
Systems Unified
Seven legacy platforms connected via single integration layer
60
0%
Exception Response Time ↓
Coordinator response time from alert to resolution
✦
Cargo visibility went from 4–24 hours to real-time — the integration layer collapsed the latency that had defined Crowley's operational reality for decades. Any cargo manager, anywhere, now has a live view of every shipment at any stage of its journey.
✦
The coordinator team's scope expanded, not contracted — freed from information relay, coordinators moved into exception resolution and proactive customer management. Headcount was preserved while output per coordinator increased by over 3x.
✦
Customer satisfaction scores increased 31 points in the first six months — self-service cargo tracking eliminated the most common customer service contact reason. Customers who previously called for status updates now check themselves. The relationship shifted from transactional to strategic.
✦
The predictive layer flagged 73% of delays before they became customer-impacting incidents — the shift from reactive to predictive operations is the structural change that compounds over time. Every delay flagged 48 hours early is a delay that gets managed rather than survived.
"

We've been moving cargo since 1892. We've survived container shipping, globalisation, and COVID. But we'd never actually been able to see the whole operation in real time — not once in our history. The Maritime OS changed that. The 47% efficiency number is real but it understates the change. What Raghvendra built is not a dashboard. It's an operational brain we've never had before.

VP Operations
Crowley Maritime · 2022
07
Section 05
Key
Learnings.
What legacy modernisation taught — and what it permanently changed about how enterprise product work gets approached.
01
01
Legacy Systems Contain Intelligence.

Every legacy system is a crystallised record of decisions made by smart people responding to real constraints. The instinct to replace it wholesale destroys institutional intelligence that took decades to accumulate. The correct approach is to understand what the system knows, preserve what it knows correctly, and improve what it knows wrongly. Replace as a last resort.

02
02
The Human Layer is Architecture.

The coordinator team at Crowley was not a bug in the system — they were a feature. They held the organisation's ability to handle exceptions, maintain customer relationships, and adapt to events no system had anticipated. The modernisation succeeded because it made that team more capable, not because it tried to automate them away.

03
03
Integration Before Interface.

The Maritime OS's most important design decision was not a screen — it was the canonical cargo event model. Without a shared data layer, every interface is a different view into a different truth. Spending eight weeks on data architecture before opening a design file was the investment that made everything else possible.

04
04
Reactive to Predictive is the Transition.

The biggest value unlock in enterprise operations is not better visibility into what's happening — it's early warning of what's about to happen. The predictive layer was technically the last thing built, but strategically it was the most important. Shifting Crowley from reactive to predictive operations changed the nature of the coordinator's job — from firefighter to architect.

← Previous Case Study
Financial
Inclusion
Rapipay · Fintech · $3B+ GTV
←
Next Case Study →
Commute
Architect
Shuttl · Urban Mobility · 1M+ Daily Rides
↗

Legacy System
Weighing You Down?

Product Strategy Sprint is the engagement for organisations with operational infrastructure that has outgrown itself. 8 weeks, transparent pricing from ₹5L — from audit to architecture to roadmap.

Initiate Inquiry ↗

Explore More
Work.

Seven more case studies across design systems, fintech, AI strategy, design leadership, and product — all with the same structural approach applied here.

View Selected Work ↗