The Opportunity: A Platform Built for What’s Next
The client had a clear vision for the future of maritime workforce management. Their strategy called for a modern, cloud-native SaaS platform designed to replace their legacy system and set a new standard for operational connectivity, compliance, and crew lifecycle management.
Working with KMS Technology, the client had already invested in a rigorous strategy phase: defining product architecture, business requirements, technical designs, and high-fidelity designs. The roadmap was detailed, validated, and ready. The question was how to bring it to life quickly, efficiently, and in a way that would generate early momentum with customers and investors.
The client faced a fork in the road. Pursue a traditional multi-phase build spanning many months and significant upfront investment, or find a smarter path to early validation of the vision.
The Approach: Designing a New Model for Product Development
The client and KMS saw an opportunity to do something genuinely new: run a structured, time-boxed experiment to determine whether agentic AI development could dramatically compress the path from product vision to working software.
The model was unproven, but the goals and endpoint were clear. The clarity of the vision and willingness to innovate were precisely what made it possible to move at an unprecedented pace.
The experiment was designed with intention. KMS brought together a lean team of specialists: an agentic engineering lead with hands-on experience, the cloud architect from the original strategy phase, and a focused development squad. Together, they operated against a six-week clock to build as much of the new platform’s foundation as possible, then stop and assess.
1. Foundations First
The team began by establishing the engineering environment: Azure cloud infrastructure, GitHub repositories and CI/CD pipelines, and AI-native development workflows using Claude Code. Critically, the detailed requirements and design compositions from the client’s strategy phase were converted into structured markdown files and loaded into a project knowledge base that agents could reference directly. This gave the AI the context it needed to build with intent, not just generate code.
2. Agentic Development in Practice
With the environment in place, the team used agentic AI workflows to drive the scaffold build, using the client’s business requirements as guardrails rather than simply writing code with AI assistance. Developers operated as autonomous decision-makers within their modules, empowered to move fast.
The result was a development pace that redefined what a small team could accomplish. Over six weeks, the squad produced over 232,000 lines of code, 287 lines per hour, compared to a traditional estimate of 500 lines per day. A 4–6 times code production multiplier.
The Outcome: From Vision to Working Product in Six Weeks
At the end of the engagement, the client had something tangible, shareable, and impressive: a working version of the platform ready to demonstrate to customers, leadership, and investors.
The team completed 35–45% of the total MVP scope, work that had originally been estimated at 6,000–8,000 hours, in just 809 total team hours. The agentic squad delivered an 8–10x feature completion multiplier compared to a traditional development team of similar size.
What Was Built
Crew Hub: Full crew profile management including contacts, competencies, compliance, employment history, and availability
Vessel Hub: Vessel profiles, departments and positions, manning plans, compliance documentation, and cabin management
Ports & Routes: Port creation, route planning, and vessel assignment
Crew Scheduling Hub: Vessel-centric rotation planning with Gantt and calendar views, drag-and-drop crew assignment, and cabin allocation
Compliance Alerts: Automated identification of crew compliance gaps
Fleet Dashboard: Roll-up metrics and alerts for fleet-wide visibility
System Administration: Tenant management, role-based permissions, and configuration controls
Production Infrastructure: Full Azure cloud deployment with Infrastructure-as-Code, mature CI/CD pipeline, and tiered environments ready for migration
More than 5,400 unit tests were written agentically, providing a foundation for the next phase of dedicated quality assurance as the platform continues to develop.
What This Proves
The client’s willingness to challenge convention produced not just a working product, but a proof of concept for an entirely new model of software delivery.
The engagement reinforced a critical insight: in fast-moving markets, the ability to rapidly validate a product direction can be just as strategically valuable as long-term feature delivery. The client now enters the next phase of platform development with a platform infrastructure foundation, real and demonstrable product momentum, and the confidence that comes from having seen the model work.
For the maritime industry and for software companies navigating the tension between innovation and investment, this approach offers a replicable blueprint.