How it works

Like your best dispatcher — tireless, and on the record.

No black box. 4T4 follows the same loop a good operator does: connect, read the situation, weigh the trade-offs, propose the move — and learn from your decision. Here's exactly how.

The 4T4 loop

Five steps, running all day.

⇄
01 · Connect
Plugs in
To your TMS, GPS, email, chat and port systems. No replacement.
◉
02 · Read
Sees the situation
Turns messy signals — even a WhatsApp — into one clear, live picture.
⚖
03 · Weigh
Weighs trade-offs
Margin, hours, windows, risk — balanced against your rules, at once.
✓
04 · You decide
Proposes; you approve
The move, the reason, the risk, an alternative. Nothing runs on its own.
↻
05 · Learn
Gets sharper
Your approve / edit / reject teaches it — and the know-how stays yours.

↻ The loop repeats on every event — so 4T4 starts useful on day one and keeps getting better with use.

What's inside

It reads your world — then decides with an engine, not a guess.

The chat you see is just the surface. Underneath, 4T4 turns your real operational data into decisions using your own rules and history.

What it reads
TMSOrders, trips, documents
GPSPositions & driver hours
@ / WAEmail & chat requests
PORTTerminal & container status
→
How it decides
The decision engine
  • Applies your rules — the constraints your team really uses
  • Balances several goals at once — margin, timing, saturation, risk
  • Works with incomplete, changing information — like real operations
  • Explains why, every time — reason, risk, alternative
Why a generic AI agent isn't enough

A chatbot can talk about your shipments. It can't reliably decide them.

× A generic AI agent
  • Chats, summarises and explains
  • Guesses from general text — not your operation
  • Wobbles when goals conflict or data is missing
  • Forgets — nothing compounds into your IP
✓ 4T4
  • Decides and prepares the actual move
  • Runs on your live data, rules and history
  • Built to balance competing goals under uncertainty
  • Learns from every decision — the know-how stays yours

The defensible value isn't "just an agent." It's your rules, your data and your decision history — compounding into capability nobody can copy.

Under the hoodAgents (LLM) for chat, interpretation & explanation
 A private decision dataspace — your constraints & logs
 A multi-objective, robust, probabilistic planning engine
 A learning loop over real operator feedback
Human-in-control & learning

Your team stays in charge. The system gets smarter.

Learning, the honest way

4T4 doesn't try to capture everything upfront. It proposes a motivated move on the evidence it has — and learns from what you choose.

Every approve, edit or reject teaches it your operation. Knowledge is built through real usage — not static rule manuals or endless interviews.

You disagree? That's valuable.

When an operator overrides a suggestion, that's not a failure — it's the most valuable signal there is.

Your judgement turns into rules the whole company owns. 4T4 doesn't replace the dispatcher — it makes their judgement scalable.

Book a technical walkthrough → See how it connects →