© Thomas Lukassek

UPS is in the process of fine-tuning its management of deliveries by leveraging artificial intelligence (AI) to inject more live data into its operations.

At the end of January the integrator began the roll-out of its On-Road Integrated Optimization and Navigation platform, which gives drivers the most efficient routings, recalculating individual package delivery routes throughout the day, taking in changes in traffic conditions, pick-up commitments and delivery orders.

In addition to more accurate delivery time estimates, it will enable new products and services that require higher levels of accuracy, reliability and responsiveness, according to UPS, and the enhanced routing tool will further reduce miles driven, fuel consumed and carbon emissions.

The dynamic platform uses advanced algorithms, AI and machine learning.

Some shippers have already begun to use AI and machine learning for forward-looking computations. Faced with uncertainty, shortening product cycles and rapidly changing consumption trends, they are trying to respond to new trends by using demand-sensing tools.

According to ToolsGroup, a provider of supply chain planning tools software, demand-sensing “closes the gap between your plan and what’s actually happening in your supply chain”, by spotting changes and patterns in downstream activities.

Going beyond historical data, using IT to detect trends can significantly improve forecasting accuracy, ToolsGroup claims. This can help manufacturers adjust production output and improve inventory management.

One customer, Amplifon, a provider of hearing solutions, has seen an 18% reduction of inventory and cut obsolescence by 50%, according to the software firm.

“Demand-sensing is proven to help our customers cut short-term forecast error by up to 50% and raise inventory performance by up to 20%, while hitting or exceeding target service levels,” ToolsGroup claimed.

Hybrid demand-sensing systems use multiple forecasting techniques and data types, which range from promotional campaigns and competitors’ activities to updated sales figures, weather forecasts and social media sentiment.

Cathy Morrow Roberson, founder and head analyst of Logistics Trends & Insights, said bots which feed off social media could enter ample data into the system. UPS employs bots for its help desk and customer services, and all the data handled by these are fed into the system, she added.

POS data are highly prized for demand-sensing, but it is possible to go ahead without it, according to E2 Open, a provider of intelligent applications, which argues that all that is needed to get started is the demand forecast and data on orders and shipments.

“More than three-quarters of demand-sensing implementations use just these three internal signals, and on average they achieve a 40% reduction in forecast error,” claims E2 Open on its website.

It stresses that demand-sensing does not bring about a fundamental change in operations, or replace longer-term forecasts, but improves visibility and provides a more granular view. Hence, it can also be applied to long-term forecasting.

Air Canada Cargo applied AI to examine patterns of booking cancellations and changes in the final week prior to a flight’s departure, and found that cancellations and reductions in booked volume occur within that period at 60% of the rate of bookings. Adjusting for this can improve load factors by 6-8%.

Not surprisingly, much of the communication about demand-sensing is emanating from software providers. Shippers and logistics firms have been less vocal about their trials. This suggests that, at this point, there is still more hype than action – much of it probably fuelled by the fact that a lot of venture capital is looking for investment opportunities in logistics technology.

Ms Roberson says there is much venture capital circling over logistics, but the fact that UPS is moving forward with its use of AI indicates that the concept is gaining traction.

“If UPS is using it, there must be something to it. It is a conservative company,” she said.

Comment on this article

You must be logged in to post a comment.