Azimut Ai
This project was done as part of my work at "Triolla"
When AI-powered intelligence meets maritime security.
Turning any camera into a smart visual radar for ports and coastal defense organizations.
Process
Research-first, then two competing directions
1. Research
Reviewed the existing platform, mapped user needs through stakeholder sessions, and studied competing maritime monitoring and command-center tools to identify what to keep and what to rethink.
2. UX Concepts
Explored multiple structural approaches focusing on the live map view and the vessel inspection flow, then handed wireframes to the client for real user testing with port operators.
3. Design Concepts
Developed two contrasting styles to explore different approaches: a modern high-tech look with glassmorphism, versus a vibrant light mode with 3D icons.
4. Refinement
The chosen concept was refined and systematically extended across all screens, states, and edge cases to complete the platform design.
Challenge
Designing clarity inside complexity.
The main challenge was building a real-time monitoring interface for operators who spend up to 95% of their time passively watching and need to act decisively in seconds when something appears.
Existing systems relied entirely on post-event footage review. There was no live anomaly detection, no alert logic, and no way to understand normal maritime behavior across large coastal zones. The platform needed to change all three. At the same time, the Visual Recognition feature, which shows AI confidence scores per angle of a detected vessel, needed to be compelling enough to anchor investor demos and client pitches.
Design
Turning raw AI and mapping data into intuitive operator workflows
Dark, high-contrast UI keeping operators immersed in the feed while critical data surfaces only when needed. Color and hierarchy do the heavy lifting, so operators don't have to. The Visual Recognition panel was one of the most considered moments in the design. The solution, a 3D frame of vessel in real time, comparing to an archive photo of the vessel, rendering with per-angle confidence overlays, turns a raw AI output into something immediately readable, even for non-technical stakeholders in a demo context.
