Eastern Pacific Shipping is expanding its use of Orca AI's maritime operations platform to dozens of additional vessels following a pilot that produced a 48% reduction in close encounter events per 1,000 nautical miles and a 10% increase in average minimum passing distance.
The five-vessel pilot ran long enough to generate data that EPS judged reliable for a fleet-scale decision. One vessel, the Pacific Sentinel, recorded an 84% reduction in open-water close-encounter events and a 50% increase in minimum passing distance within six months of installation — the kind of result that tends to accelerate deployment timelines.
Orca AI's platform has two components. SeaPod sits on the bridge and uses computer vision to detect and classify vessels and objects in the surrounding water, including targets that do not show up on AIS (the standard maritime tracking system) — fishing boats, buoys, and other hazards that traditional navigation aids miss. FleetView handles the shore side, giving operations teams access to navigational event logs, trend data, and the material needed for data-driven crew training.
That combination of real-time onboard awareness and retrospective shore-based analysis is increasingly common in maritime AI deployments, as operators look for technology that improves safety without requiring bridge crews to change their core workflows.
EPS's technical director described the pilot methodology in straightforward terms: objective data collection on how navigational risk changed across a defined period, rather than a subjective assessment of the technology. That framing matters because maritime AI vendors have historically had to work against scepticism about whether onboard computer vision systems can perform reliably enough in live conditions — in congested ports, in low visibility, and at the speeds and distances where close encounters turn into incidents.
The Pacific Sentinel numbers, if representative, suggest the platform's performance in real operations aligns with what Orca AI publishes in its marketing materials. The expansion to the broader EPS fleet will be a more demanding test, operating across varied routes and conditions rather than a structured pilot.
To stay across the latest in cloud, AI and enterprise tech analysis from Compare the Cloud, subscribe to our weekly newsletter at https://www.comparethecloud.net/newsletter
