AI-powered vessel operations for port logistics.
How a ship chandler at the Port of Tampa replaced hours of daily inbox triage with an AI agent that reads emails, reasons over operational context, and proposes structured updates — with a human always in the loop.
Dozens of unstructured emails. Zero structure.
A ship chandler at the Port of Tampa supplies roughly 40 recurring vessels with provisions, fuel, and services. Every day brings dozens of unstructured emails from captains, agents, and port authorities — ETA changes, customs holds, berth reassignments, voyage bookings, cargo updates.
Each one buried in casual prose, forwarded threads, automated notifications, and one-off formats. Staying on top of which ship is where, when, and what it needs becomes a full-time triage job that pulls the operator away from actually running the business.
A complete operational system.
Not just an AI that reads emails. A full operational layer that replaces the mental model the operator was carrying around in his head — which vessels are active, what they need, when they're arriving, and what changed since yesterday.
An AI agent ingests every inbound email, identifies the vessel, locates the active voyage, and extracts what changed — new ETA, berth reassignment, customs hold, cargo update. It proposes structured updates that the operator reviews and approves with one click. Nothing touches the operational record without human sign-off.
The result is a living, trusted dashboard — a single source of truth for every vessel, every voyage, every supply need. Built on tools the operator already uses, with an AI reasoning engine doing the parsing, matching, and structuring that used to eat hours of his day.
The operator went from managing his business out of an inbox to managing it from a single queue he can actually trust.
Five steps. One review queue.
Emails from captains, agents, and port authorities route to a dedicated inbox. Each one lands as a raw record in the system — nothing gets lost, nothing gets skipped.
An AI agent identifies which of the 40 vessels the email concerns — by name, IMO number, call sign, or alias. It locates the active voyage and extracts proposed changes: new ETA, status shift, berth reassignment, cargo update, customs notes.
Rather than silently updating operational data, the agent creates a staging record that previews the change side-by-side with current values. Current vs. proposed, with the source email attached.
The operator sees a clean review queue: source email on one side, proposed diff on the other. Approve, edit, or reject with one click. Nothing changes without human sign-off.
Approved proposals push to the operational record with a full audit trail. Rejections feed a corrections loop that tunes the agent’s reasoning over time — learning vessel-specific patterns, captain habits, recurring berths.
Trust is built in.
- 01Inbox triage drops from hours to minutes.
- 02Operational data stays accurate and timestamped.
- 03Nothing changes without human sign-off — trust is built in by design.
- 04The agent gets smarter with use, learning vessel-specific patterns without anyone writing new code.
One queue you can trust.
The inbox didn't go away. The chaos did.
Before: the operator dug through dozens of email threads every morning trying to piece together which vessel was arriving when, what it needed, and what had changed overnight. Critical updates got buried. Things slipped.
After: a single, automated queue — fed by those same emails — where every vessel's status, ETA, and supply needs are current, structured, and trustworthy. No digging. No guessing. No “wait, when was that boat coming again?”
Same data, same emails — transformed from noise into the single source of truth the business runs on.