# Navatom AI — AI for Ship Management
Ask your fleet anything. Navatom AI is an assistant built into your ship-management platform. It answers in plain language from your own live crew, maintenance, procurement and HSEQ data — grounded in your manuals and the regulations, and it always shows its work.

## At a glance
- 6: Checks in your morning brief. Approvals, overdue and new issues, certificates, crew requests and safety reports.
- Cited: Every answer shows its source. The record, the manual section, the circular or the regulation.
- 0: Changes without a person. Every change is a draft that someone approves.
- Ship + office: The same assistant on board. With a data saver for slow satellite links.

## How Navatom AI works
1. Ask in plain language: From any screen. The assistant knows what you are looking at, so you can ask in the words you already use.
2. It grounds and cites: Pulling from your manuals, the regulations, and your live records — with links back to the exact source.
3. You stay in control: For any change, it drafts a proposal you review and confirm. Nothing is ever written on its own.

## What it does
- Your day, briefed before you ask: Open Navatom AI and the Today brief is waiting: the approvals, overdue issues, new issues, expiring certificates, crew requests and safety reports that need you, across the vessels you look after. Ask what to focus on first.
- Answers you can take into the meeting: Ask for a list and get a table, a chart when numbers need comparing, and a PDF or Excel export when it goes to the meeting. Share the chat as a read-only copy, or add it to the record as a note.
- The AI proposes, you approve: Ask for corrective and preventive actions, an investigation report or a clearer description, and Navatom AI drafts them for you to check. A change that is hard to undo waits until you tick it, and nothing is saved until a person applies it.
- On board, even on a slow link: On board, answers stay short and photos small while the link is slow. Ask while the satellite is down and the question waits, then goes when the link is back. A draft you create on board is saved on board and reaches the office with the next sync.
- It keeps working when you step away: Reports that arrive on schedule, alerts that watch for you, playbooks your team runs with one command, and a memory of how you work.
- Built to be trusted with your fleet: It only sees your company’s data, and only what the person asking could already open in Navatom, vessel by vessel and module by module. Every interaction is logged for review, and your data never trains a model.
- Quick, Balanced, or Deep: No models to compare, no vendors to weigh up. Choose how much depth the question in front of you deserves: Quick for everyday lookups, Balanced for most work, Deep for analysis that needs time.

## While you are away
- Reports and alerts on a schedule: Ask once and a report arrives every weekday morning, or an alert checks every 15 minutes and tells you when a critical job falls overdue.
- Playbooks for the questions you repeat: Save the instructions you give again and again, then start them with a / command. Share them with a person, a user group or a vessel.
- A memory of how you work: If you allow it, Navatom AI keeps short notes about your work and preferences and uses them in new chats. Only you can see them.

## What Navatom AI will not do
- It never changes your records on its own. Every edit is a draft you review, adjust, and confirm. Nothing is written until a person signs it off.
- It’s not a black box. Answers cite the manual, circular, or record they came from — or say plainly when they don’t know.
- It’s isolated to your company. It only ever sees your data, filtered by each user’s permissions — and it’s never used to train a model.
- It stays in its lane. It reports what’s in your system. It doesn’t invent numbers, forecast failures, or track vessels live.

## FAQ
### Does Navatom AI make changes on its own?
No. Navatom AI only drafts. Every edit — a corrective action, an updated description, an investigation report — is a proposal you review, adjust, and confirm. Nothing is written to your records until a person signs it off.

### Which data can Navatom AI see?
Only your company’s data, and only what the signed-in user is already allowed to see. It is tenant-isolated and permission-aware: every answer respects each user’s vessel access and role, and every interaction is logged for review.

### Do you train models on our data?
No. Your data is never used to train a model. Navatom AI runs on enterprise-grade foundation models with EU data residency, and it answers by retrieving from your own documents and records — not by memorising them into a model.

### Can it predict equipment failures?
No. For planned maintenance it reports what is overdue and the running-hours and downtime history already logged in your system. It does not forecast failures or invent maintenance dates.

### Where do its answers come from?
From your manuals, the regulations, and your live records. Every answer cites the source it used — the manual section, the circular, or the record — or tells you plainly when it does not know.

### Do I have to choose an AI model?
Never. You pick one of three settings — Quick, Balanced, or Deep — depending on how much depth a question needs. There is no model jargon and no cloud vendor to choose.

## Links
- Page: https://navatom.com/solutions/maritime-ai
- Start a free trial: https://app.navatom.com/trial
- Talk to sales or book a demo: https://navatom.com/contact