Transportation and logistics provider XPO Logistics Inc. will use four software initiatives to optimize the flow of its less-than-truckload (LTL) transportation network in a push to generate a $100 million incremental increase in operating profit within two years, the company said Thursday.
Greenwich, Conn.-based XPO, which is the second-largest LTL provider in North America by revenue—trailing FedEx Freight and leading Old Dominion—calls the initiatives the core of its "LTL 2.0 optimization plan." The initiatives rely on computer algorithms and artificial intelligence (AI) to squeeze inefficiencies out of the firm's routes, loads, and prices.
Two of the innovations are already operating in pilot stage, according to a statement from XPO chief information officer Mario Harik. They include a linehaul bypass model that uses algorithms to assemble truckloads dedicated to direct movements of LTL freight, helping trucks avoid making multiple stops at a single destination.
Also in pilot mode is an AI tool that supports dynamic route optimization for pickup and delivery by setting the sequence of each LTL driver's pickups and deliveries, and then adjusting those routes based on traffic, road construction, and other real-time variables.
Next in XPO's technology cue will be a load-building application that applies AI tools and computer vision sensors to help cross-dock operators ensure that inbound pallets are loaded in the optimal trailers and in the correct sequence to boost delivery efficiency and limit damage.
Finally, XPO is developing pricing algorithms that will use machine learning capabilities to predict LTL price elasticity, allowing users to forecast market conditions for future dates based on current supply and demand, historical data, and XPO's capacity levels.
XPO says it supports the technology development through a $450 million research and development budget that has also supported projects in warehouse robotics, voice-operated interfaces, and a digital freight marketplace in the firm's mobile app for drivers.