Fujitsu and partners trial AI crowdsourced delivery app

Image credit: Andrey Suslov /

Fujitsu, Singapore Management University (SMU), A*STAR’s Institute for Infocomm Research (I2R) and UrbanFox (a subsidiary of Keppel Logistics) have signed an agreement to collaborate on a field trial in Singapore in which they will test a smartphone app that offers AI-driven delivery recommendations to crowdsourced delivery personnel.

The trial solution is intended to address the growing demand for deliveries from e-commerce purchases in Singapore. At 5.4%, the country has the highest online share of retail sales among Southeast Asia’s five top economies, and the e-commerce trend shows no signs of slowing down anytime soon.

UrbanFox’s solutions suite includes dynamic logistics support powered by a crowdsourcing model for last-mile deliveries. The company currently handles thousands of deliveries per day, and the number of deliveries more than double during peak-season events like Singles’ Day. As delivery volumes continue to grow in tandem with e-commerce, smarter solutions are needed for its crowdsourced delivery personnel to choose optimal delivery jobs from the huge volume of delivery orders available, taking into consideration such factors as efficient delivery routing.

With the Asian e-commerce market predicted to expand rapidly in the next five years – and with Singapore experiencing manpower shortages in the logistics industry – one potential solution will be to utilize AI to help keep pace with the expected increase in delivery orders. Fujitsu, SMU and A*STAR’s I2R will collaborate closely with UrbanFox on a joint trial that will test and evaluate technologies intended to optimize the assignment of delivery jobs to each delivery partner.

Specifically, the trial will incorporate AI into a system that recommends delivery jobs and delivery routes optimized to delivery partners with the goal of improving the productivity of delivery tasks.

During the test bedding phase, which begins this month, the organizations will analyze order data managed by UrbanFox, such as the geolocation data of delivery partners as well as their past delivery performance. This data will then be matched with the requirements for delivery, and a recommendation of the most efficient delivery partner will be given. Ultimately, about 30 delivery partners are expected to participate in real world deliveries over the course of the trial.

Delivery partners can choose whether to accept the recommended delivery assignments – the organizations will conduct machine learning on those decisions, continually improving the accuracy of recommendations.

fujitsu AI crowdsourced delivery
Image credit: Fujitsu

I2R will incorporate its proprietary AI-enabled descriptive and predictive analysis algorithms into the system to optimize the delivery process. By leveraging AI, the system will also be able to provide insights into delivery trends for an area. This results in the system being able to provide predictions on delivery demands for an area based on past deliveries and events such as sales.

When an order is received, the algorithm will first review the delivery requirements, such as size of the item and delivery route. At the same time, a trade-off analysis will also be performed on the data of delivery partners with UrbanFox to determine if it is more efficient to use a delivery partner, or the company’s own delivery fleet. This will help optimize the delivery process.

SMU will conduct research on an AI-based recommendation approach to automatically suggest bundles of delivery tasks that are most suitable for each delivery partners. These recommendations will be personalized and dynamic, reflecting personal preferences and real-time status of both the delivery personnel and the delivery demands.

Dispatch planning technology developed by Fujitsu will be subsequently leveraged to set delivery plans and calculate efficient delivery routes, notifying delivery partners of the recommendations through this app.

Be the first to comment

What do you think?

This site uses Akismet to reduce spam. Learn how your comment data is processed.