How Transwin Made Mirage Its AI Agent Partner

About Transwin

Transwin is a Moroccan logistics operator built around port flows. Founded in 2015 and operating from El Jadida, Jorf Lasfar and Casablanca, it moves agri-food, fertilizers, steel and mining freight with a fleet of 550 specialized trucks and 100 machines, across 40,000 m² of warehousing. The company has been growing at roughly 50% a year.

That growth is what created the problem. Operational data lived in separate systems, spreadsheets and teams. Adding trucks and contracts meant adding manual work, and no one had a single view of what was happening across the business.

Forward-deployed, not off-the-shelf

Every system below was built the same way: Mirage engineers on site, working from Transwin's actual documents, trucks, tanks and TMS, rather than a generic platform Transwin had to adapt to. That's the approach that turned one paid document workflow into five production systems across freight, fuel, mining and fleet, each new domain got the same embedded build process, not a resold module.

Where it started: one workflow

Transwin's operations teams were reading freight and customs documents by hand before keying the data into their TMS. Mirage deployed its logistics agent, built against Transwin's actual document mix and wired directly into their TMS. The scope started with transport documents, then grew to include invoicing and financial documents as well.

The same agent then took on inbound tenders. It processes both emails and scanned physical documents, extracting and structuring the relevant information.

Then it expanded

Fuel intelligence

The fuel system (sold until September 2026 as FuelMetrics) gives Transwin real-time visibility across fuel deliveries, on-site tanks, pumps, and vehicle transactions. The platform helps teams track stock movements, reconcile delivered versus received volumes, monitor consumption, investigate anomalies, and analyze fuel performance by site and vehicle. Following the initial deployment, it is now being extended to additional sites and mobile tanks. The reconciliation method it applies is described in the fuel reconciliation guide.

Mining equipment

Transwin's mining machines fail expensively and unpredictably. As part of a second control room project focused on the mining side of the business, Mirage built per-machine tracking of greasing, oil changes and mileage, then layered prediction on top: maintenance is scheduled before breakdown, and usage data feeds how machines are assigned and routed.

Control room

A single operational view across the entire business, not just mining. Every trip is tracked in real time through a data architecture that combines telematics hardware with AI on top: live fleet tracking, route optimization, speeding alerts on drivers, and consolidated KPIs. The system surfaces anomalies and events that need a human.

Private assistant

Before the LLM came the data platform: Mirage first unified Transwin's documents and operational records into a single structured base. The custom LLM is grounded in that platform, answering questions about Transwin's business rather than the world.

What changed

Transwin started with one document workflow and a paid deployment. Mirage now operates across four areas of the business, built specifically for how Transwin runs.

Where it goes next

› Predictive operations
› Mining production monitoring
› Executive dashboards
› Automated operational reporting

Questions about this deployment

• What does Mirage run at Transwin?

• What are the measured results?

• Who is Transwin?

• How did the deployment grow?