The transport industry is the heart of commerce and the linchpin of a functional economy.
Moving goods more efficiently, more effectively and at the lowest cost possible isn’t going out of style anytime soon.
Now there’s a chance for artificial intelligence to play a role in the sector.
“I’ve been in the artificial intelligence and machine learning space for more than 20 years now,” Yoav Amiel, chief information officer at freight brokerage platform and third-party logistics company RXO, told PYMNTS during a conversation for the “AI Effect” series. “I’ve seen the evolution of the field. And there are three areas where today’s AI capabilities differ from previous predictive machine learning solutions.”
For one, Amiel explained, the ease of building AI models has improved with the availability of AI platforms and open-source tools.
“To build generic AI capabilities in today’s world is a lot easier,” he said. “The cloud has driven a lot of capabilities … open-source engineers can take a snippet of code and produce a new AI engine relatively easily.”
The ability to more easily build AI systems has in turn led to AI algorithms becoming more complex, providing a greater depth of possibility and context around the results AI can generate and derive from information.
Lastly, Amiel said the adoption of AI has increased thanks to a more widespread availability of training datasets and other necessary technical resources.
“It has democratized AI development,” he said.
What that means is that the stage is set for AI to comprehensively evolve transportation sector workflows.
While the text-based capabilities of generative AI systems, like OpenAI’s ChatGPT product, have captured public imagination, the enterprise use of AI’s multimodal capabilities, including across video and imagery, represent some of the innovation’s most promising applications.
“If you look at the maturity of AI models over the years, if you go back 20 years, AI was more around recognition, and gradually that evolved into coming up with insights and serving as a recommendation engine,” Amiel said. “Today, AI is capable of task completion — and that’s what gets me excited. If you think about warehouse inventory planning, workforce planning, all of these activities, AI can make these processes much more efficient overall.”
Within the transport sector, multimodal AI systems capable of dynamic task completion can optimize pricing, scheduling and routing in real time, leading to increased efficiency, he explained.
“When we build technology, we’re not building it just for its own sake,” he said. “We build technology to help the business, and there are three key business levers: productivity, margins and volume.”
He emphasized that effective deployment of AI can drive a positive impact across each. The offline and highly intermediated nature of transport and logistics is ripe for data-driven transformations that make the sector a more connected and predictable ecosystem.
When asked about the potential for AI to enhance cross-border operations, Amiel highlighted applications such as dynamic route adjustments, multilingual chatbots and document processing.
While AI offers benefits, Amiel acknowledged the need to address potential issues including AI’s safety, reliability and auditability, which he listed as crucial factors to ensure trust in AI systems. The ability to handle unforeseen scenarios, minimize algorithm bias, and protect privacy and cybersecurity are also important considerations, and Amiel stressed the importance of testing, certification and transparency in AI systems to maintain fairness and accountability.
“Another area to be aware of is the dependency on technology overall,” he said. “We are giving more and more decision-making power to technology, and we need to make sure that if the machines are for some reason unable to make these decisions, we are not left without the ability to function. With applications of AI within self-driving cars and trucks, this is becoming more important.”
After all, the adoption of AI is still in its early innings. While there may be a learning curve, AI systems are becoming more intuitive, making adoption more accessible to a wider range of users, as well as underscoring the risk of AI systems becoming single points of operational failure.
Still, Amiel said that the transportation is ready — and willing — for AI adoption. Larger companies with more resources have been at the forefront of implementing AI solutions, but the democratization of AI is gradually eliminating barriers for smaller companies.
“Long-haul trucking is much more open to automation,” he said. “There are more potential savings and efficiency gains there that are easier to implement. Short-haul trucking has higher implementation costs and tighter margins, so deploying AI and embedding these engines could be more complex.”
Looking ahead, Amiel highlighted two areas of innovation in the transportation industry.
The first is task completion, where AI can automate end-to-end activities, from planning to delivery. This can streamline processes for shippers, carriers and internal operators.
The second area is smart infrastructure, leveraging the Internet of Things to optimize traffic management, accident prevention and routing. Dynamic routing algorithms can consider factors such as traffic, vehicle availability and delivery deadlines.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.