Orbbia OSMarine Operations Intelligence
PT Talk to us

Marine Operations Intelligence

An AI agent layer on top of the maintenance system you already run.

It tells you which vessel needs a decision this week, before the class window and the cash ceiling collide.

AI recommends. Your team approves.

See the demo How it works

01The problem

The rule is fixed and public. The decision is still built by hand.

01

The most expensive event in a vessel's life

Drydocking is the largest cost event of the life cycle. The rule that forces it is public: a Special Survey every 5 years and a drydock at most every 36 months. Nothing about the date is a surprise.

02

The PMS shows the state. It says nothing about the decision.

A planned maintenance system records tasks, hours, inventory and certificates. It tells you how the vessel is when someone goes to look. It does not say which vessel needs a decision this week, or what that decision costs.

03

Class window and cash ceiling collide.

Two drydockings land in the same quarter as the fleet's lowest cash. Somebody has to see that six months ahead. Today that somebody works with a spreadsheet, an inbox and memory.

Sources for the rule: IACS UR Z3 and Z7, ABS Rules Part 7, NORMAM-201.

02How it works

Three steps. One approval gate.

  1. 01

    Read

    Orbbia OS reads what your maintenance system already knows: tasks, hours, certificates, class dates. And the public record of your fleet: class status, survey windows, age.

  2. 02

    Recommend

    Agents cross the class window with the cash ceiling and the yard calendar. Each recommendation arrives as one card: why, confidence, action, and who approves.

  3. 03

    Approve

    A named person on your team approves, changes or rejects. Nothing moves without that signature. The log keeps every decision.

  4. Approval gate

    AI recommends. Your team approves. Never the other way around.

03Modules

Three modules, one operating model.

01 · Drydock Intelligence

Drydocking planned as a project.

How do we run each drydocking as a project, from D-180 to the purchase order?

  • The project opens itself 180 days before the yard, with the class window as the deadline.
  • Scope, budget and yard slot decided in sequence, each with its approving authority.
  • Buy as a fleet, not as a vessel: repeated items grouped across dockings.

02 · Readiness and Port Call

Is this vessel ready?

Is this vessel ready, and what has to happen at the next port call?

  • Certificates, surveys and open work orders read as one readiness picture.
  • What must be closed before the next port call, and who owns it.
  • An alert when a certificate or class date drifts into the operating window.

03 · Predictive Maintenance

What fails, before it fails.

What is going to fail, before it fails?

  • Failure patterns read from the maintenance history of the fleet, not of one vessel.
  • A recommendation, not an alarm: why, confidence, and the action to take.
  • Direction, not invented percentages: the agent says what the data supports.

Screens from the generic demo. Fictional fleet, illustrative data.

04Demo

See it working, with a fleet on screen.

The generic demo runs a fictional fleet with illustrative data, end to end: fleet, drydock projects, constraints, approvals, procurement and the agent's cards. Sign in with any e-mail.

Open the demo (en-US)

Your fleet in the demo, in 48 hours.

Before the first meeting we check your fleet in the public record (class, survey windows, age) and bring the demo back with your vessels on screen. No invented ROI: the calculators use your day rate.

Ask for it by e-mail

05About

Who we are

Three people who watched drydocking decided by spreadsheet, inbox and memory, from both sides of the table: the engine room and the data. Two of us come from operations, on the floor of vessel operators in the United States and in marine support in Brazil. One of us is a mechanical engineer who spent years reading operations data before building with AI. We built the demo first, then the company.

Avner Vasconcelos

Product, data and AI

Portugal

Mechanical engineer. Years reading operations data before building with AI. Builds Orbbia OS.

Hugo

Operations and commercial, United States

Virginia, USA

Engine-room background in vessel operations in the United States. The operating voice of Orbbia OS in the American market.

Leonardo

Operations and commercial, Brazil

Rio de Janeiro, Brazil

Marine support operations in Brazil. The operating voice of Orbbia OS in the Brazilian market.

How we work

AI recommends. Your team approves.

Every recommendation carries a why, a confidence level and the authority that signs it. No agent decides alone.

Direction, not invented percentages.

A number appears only with a source: a public rule, your own records or your day rate. Benchmarks are stated as orders of magnitude.

The fleet is the hero.

Orbbia OS is the layer that gives the technical director the decision back, with time to spare. It sits on top of the system you run. It does not replace it.

The long view: an AI operating system for physical operations. Marine is the first vertical.

United States · Brazil · Portugal

06Contact

Talk to us

One e-mail is enough. Tell us the size of the fleet, the maintenance system you run and when the next drydocking is due.

info@orbbiaos.com

Prefer a 30-minute strategy session? Say so in the e-mail.