Artificial Intelligence is the next big thing, logistics has always been a big thing. So why aren’t they often mentioned together?
Logistics is all around us. The clothes we wear, the food we eat, our smartphones, everything we use and consume on a daily basis has been brought to us by logistics. Take a look around you. In all likelyhood, the building you are in (or the building nearest to you) has been built with plywood coming from Brazil, cement coming from China and tools coming from Germany, which met thanks to logistics.
The enormous amount of data constantly generated by warehouses, depots, distribution centers, delivery routes, shipments and other aspects of the supply chain makes logistics the perfect field in which to apply Artificial Intelligence algorithms. And yet, save (possibly) for some of the largest players, logistics hasn’t generally been able to keep pace with the recent developments in A.I. As pointed out by a recent collaborative report by IBM and DHL:
only 10% of current systems, data, and interactions include elements of AI analysis and results¹.
The reasons behind such a tardy and lacking adoption are both historical and technical. Logistics is an an activity that has literally always existed. Large-scale civil and military logistics were already practiced as far back as Ancient Persia and Greece, while the Roman legions owed much of their might to an efficiently organized supply chain². Today’s logistics is the result of over two millennia’s worth of experimentation, trial and error, progressive additions and failed attempts. Swift innovation in such a stratified discipline is impractical.
But this is just one side of the story. Acceptance of Artificial Intelligence is also hampered by a set of more practical reasons. The modern, global marketplace has spread out the supply network and extended the reach of local manufacturers. Goods can sit in tens of different depots across several continents before they reach their final consumer, and at every step of the way the merchandise is typically handled by a distinct carrier. Before getting into your hands, the device you are using to read this article has likely been through more countries than the average person will ever visit in his or her entire life. This heterogeneous and composite system, and the fragmentary data it produces, is hardly ideal for the adoption of a technology that more often than not relies on a steady supply of clean and uniform data.
For one thing, heavy-duty introduction of Artificial Intelligence into logistics will require specialized providers. The same software makers that develop A.I. solutions for image recognition are unlikely to possess the infrastructure and know-how it takes to tackle optimal planning problems. Logistics is a very peculiar activity and necessitates specific software solutions.
Secondly, a fair degree of data standardization will be helpful. As long as each logistics firm keeps using different data formats, structures and patterns, coming up with comprehensive Artificial Intelligence models applicable to all stages of the supply chain will be a daunting task. And chances are systematizing this type of information would not prove particularly difficult: logistics companies generally have similar processes, comparable inputs and outputs and even use the same enterprise software.
Finally yet importantly, as is the case for any innovative product, incorporating A.I. into logistics will require a great deal of trust, patience and perseverance on the part of business owners, managers and workers. Much like the global supply chain we presently rely on wasn’t built in one day, Artificial Intelligence won’t become one with logistics in a few days, or a few years for that matter. Appreciable results will take time to manifest themselves, and continuous investment in research and development will be imperative, but the long-run benefits will be abundant and mutual.
The added value that Artificial Intelligence can bring to logistics is great. At its core, logistics is a large collection of organizational problems. It involves planning what to move, where, when, how and in which amount. Artificial Intelligence, which is especially efficient at recognizing patterns, estimating possible future scenarios, calculating combinations and optimizing processes, can assist logistics provider in coordinating combinatorial tasks, ensuring faster, cheaper and more efficient execution.
One of the most fitting examples of application of A.I. to the field of logistics is warehouse optimization, and more specifically picking optimization. Picking is the activity of moving through the warehouse to fetch the items that are needed to fulfill orders. Accounting on average for over 50% of all expenditures³, picking is by far the most expensive warehouse operation. Artificial Intelligence can drastically reduce picking costs and times by computing the shortest picking paths among all possible combinations.
A second example is the optimization of delivery routes. In much the same way as shortening picking paths can reduce warehouse costs, traveling shorter delivery routes can reduce shipment costs. Think of a mailperson having to deliver 100 letters to 100 different houses. What’s the sequence in which to visit all 100 houses, such that the complete path is shortest? Being able to find that sequence can decidedly increase the efficiency of the process.
Merging A.I. and logistics will be one of the grandest goals of our century. A goal which, if brought to completion, will fundamentally reshape our supply lines and lead to faster, cheaper and more sustainable logistics. This will have a profound impact on our consumption habits and on our very notion of use of goods.
Adamas is a startup which develops Artificial Intelligence software for the optimization of logistic procedures. Do you want to know more about us? Visit our website and follow us on LinkedIn!