The Future of AI in Freight Logistics

by Veronica Turk

Artificial Intelligence used to be relegated to sci-fi shows and comic books. Now, it seems like everywhere you turn there is a new company talking about AI especially in the freight and logistics industry. The most recent catalyst for this AI takeover happened in November of 2022 when OpenAI launched its initial version of ChatGPT, a large language model (LLM) that has since democratized and normalized AI usage around the world. AI is no longer a futuristic concept, it’s talked about over every kitchen table.

ChatGPT with its sophisticated underlying LLM and other competing models (e.g. Claude, Gemini, Llama) are certainly the most popular innovations. So, companies are employing these types of generative AI to help with complex information retrieval and to accelerate processes. Besides LLMs, there are other fields in the machine learning (ML) world to improve businesses, such as computer vision and robotics. While AI is having impacts across the board, no industry is as primed to benefit more from artificial intelligence in the present and near future rather than the freight and logistics industry.

How AI is being used in logistics

Currently, many supply chain companies are leveraging existing AI capabilities to power new and advanced technologies. These new technologies can plug and play in shippers’ logistics operations or be molded to meet their unique needs. Most rely on a variety of types of generative AI, but for our purposes we’ve grouped them into main categories for easier understanding. Let’s dive into some of the ways AI is being used in freight and see how AI is disrupting the future of logistics.

Large Language Models

LLMs have been designed to understand, generate, and interact with human language (e.g. GPT as a flagship example of LLM). They’re a type of deep learning model that has been trained on vast amounts of textual data, so that they can perform a wide range of natural language processing (NLP) tasks.

Chat and customer support - In the logistics industry, many technology companies are employing LLMs as an aid to their customer service and support functions. Interactive chat bots give shippers a self-serve way to get their more routine questions answered without having to contact support or read through help center sites. Some are even developing capabilities to answer more complex questions like the age-old “Where’s my truck?”

Computer Vision

Computer Vision is a field of AI that focuses on enabling computers to interpret, understand, and process visual information. In a nutshell, it’s giving computers the means to see the world around them through videos and still images. Computer vision programs rely on cameras and sensors to operate, much as humans rely on their eyes to see.

Gate and yard visibility - These days, most gates and yards are manually overseen by a human. Someone sits at the gate to check drivers in, and if you need to know whether a specific trailer is in the yard, you have to go look it up (or worse, physically look for it.) Computer vision technologies like NavTrac are eliminating these manual processes, by capturing data automatically and more reliably. Through cameras installed onsite, warehouses can automatically track containers, tractors, and chassis, enabling real-time inventory reporting and damage detection.

Shipment scanning - Computer vision enabled solutions for enhanced sorting and shipment processing are available on the market today. These technologies can accurately identify, categorize, and manage parcels and pallets, even in complex environments. This helps ensure correct tracking and routing, which significantly reduces errors and operational costs. It can also greatly improve visibility to what’s on a truck even after it’s left the yard, helping shippers maintain complete control over their shipments in-transit.

Other Machine Learning (ML) Initiatives

Machine Learning (ML) is a subfield of artificial intelligence that leverages statistical algorithms and other techniques to learn from historical data and generalize unseen data, performing tasks without explicit instructions or supervision. It then can help the user to make decisions or give predictions about the future in a very efficient and cost effective way. Over time, machine learning systems can improve their outputs by using more and varied data. The more quality data an ML technology is exposed to over time, the better the insights that can be delivered.

Freight Intelligence - One of the most impactful and complex uses of AI in the freight and logistics industry is using machine learning to interpret enormous datasets to discover and present the most significant areas for improvement in the shipment cycle. Doing so requires data inputs from disparate sources like TMSs, ERPs, WMSs, and internal data lakes.

Loadsmart’s FreightIntel AI solution is an AI-powered solution tailored to find and present insights based on shippers’ own data without the user needing to know what questions to ask! It combines shippers’ logistics data with trusted market benchmarking metrics and uses AI to generate data-backed next steps tailored to the individual shipper’s needs. It serves as a digital data analyst with deep logistics experience for any shipper to plug into their processes. And since it works by continually ingesting logistics data over time, teams can rely on it for ongoing optimization of their processes.

Predictive Maintenance - AI models consider known factors like asset age, mileage, maintenance calendar, and other data points, to recommend maintenance and avoid downtime. This removes the highly manual process of tracking asset usage and helps teams be proactive in their maintenance, reducing costs as well as saving time. Additionally, LLMs can be used to generate automated messages to staff communicating the needs of specific vehicles.

Robotics

Artificial intelligence combined with robotic systems is a burgeoning industry. It’s still mostly ‘future-state’ for the supply chain, though is expected to have a significant role to play in the future of freight. While many AI-enabled robotic functions have been researched and proven, adoption has been slower than other AI applications in the logistics industry, largely because of the capital expenditure needed to implement robotics.

Autonomous trucking - This is likely the application of AI that comes to mind when you think about AI-enabled robotics! People have been talking about autonomous trucking for years and there have even been large-scale trial runs in southwestern states. Dozens of companies are vying to have the first large-scale deployment of driverless trucks. The first company that can get it right in terms of safety, practicality, and cost will immediately jump to prominence in the industry.

As of 2022, there are some 3.54 million truck drivers in America and a constant driver shortage plagues fleet operators. It seems unlikely that other modes will replace truckload shipping anytime soon, so being able to transport more without the need for additional truck drivers is appealing for our consumer-centric economy.