French postal service company, La Poste, has partnered with Pathway, the data company that builds live AI, to create a digital twin of the La Poste logistics fleet.
Drawing on raw IoT data and sophisticated data infrastructure tools, Pathway has significantly driven the company’s ability to make highly accurate Estimated Time of Arrival (ETA) predictions. The new tool has driven IoT platform efficiency with real-time predictive analytics and cut fleet CapEx spend by over 10%. And delivered what its Head of Technological Innovation, Jean-Paul Fabre refers to as “paradigm shift” and “enormous” ROI.
Fast and accurate ETA predictions are essential for logistics firms to manage supply chain operations. La Poste’s sprawling network moves parcels through 17 industrial platforms and orchestrates more than 400 truck movements each day.
La Poste’s use of legacy infrastructure meant that the 16 million IoT-generated geolocation points collected each year from its fleet were being underutilised. As a result, La Poste’s Technological Innovation team adopted Pathway’s Rust-native streaming engine and built microservices with Delta Lake communication, from data preparation to prediction to evaluation. This enabled it to harness this untapped data without the need to overhaul its technology stack.
Estimating ETAs accurately at scale, only by leveraging the IoT data, is a complex task. It requires making sense of multiple streams of IoT data generated by transport units. La Poste utilised Pathway’s processing engine to build a data pipeline that prepares and curates the raw data required to make sub-second ETA predictions. Then, they designed sophisticated components around the initial data preparation pipeline to boost its ability to continuously evaluate and improve ETA predictions.
This framework was underpinned by three core components:
- A data cleaning pipeline to streamline, clean, and detect data anomalies directly from the raw data emitted from fleet transport units (like invalid GPS data or duplicated events producing identical timestamps)
- A ETA prediction pipeline to make ETA estimates at speed
- Individual microservices to validate actual fleet arrival times to evaluate and continuously improve the quality of ETA predictions
It effectively creates a live AI layer that acts as a digital twin of La Poste’s fleet. It transforms raw GPS points into live fleet analytics, sub-second ETAs and real-time anomaly alerts in one continuous flow.
The migration to Pathway drove optimisations across La Poste’s IoT platforms, reducing total cost of ownership by 50% and projecting a reduction in fleet CAPEX by 16%.
“Data teams working in the logistics sector need the best possible real-time processing capabilities to inform supply chain management with the most accurate ETAs. Many solutions are ill-equipped for the real-time processing of complex geospatial data, but La Poste’s use of Pathway’s data processing engine demonstrates our unique capability to inject real-time intelligence into logistics,” said Claire Nouet, COO, Pathway.
There’s plenty of other editorial on our sister site, Electronic Specifier! Or you can always join in the conversation by visiting our LinkedIn page.
