Freight logistics remains one of the most complex global systems, with countless human touchpoints creating inefficiencies and delays. While technology continues to advance, the industry still struggles with human bias, relationship-based decision-making, and communication gaps. Nilanjan Mitra Thakur has developed innovative approaches to model these interactions as intelligent agents, potentially transforming how freight moves across the globe.
Understanding Key Players in the Freight Ecosystem
The freight industry operates through two primary stakeholders: beneficial cargo owners (BCOs) and carriers. As Nilanjan explains, “If you think about the supply chain industry, whoever owns the goods and are importing or exporting their goods internationally or domestically, they are actually beneficial cargo owners or BCO’s because that’s your cargo. They’re also commonly referred to as shippers.” Meanwhile, carriers represent the transportation providers. “The carrier is mostly mentioned as an ocean carrier who actually moves the freight with ships, or it can be a trucker for domestic transport, or air carriers,” Nilanjan notes. “Basically, whoever is responsible for the modes of transportation is called a carrier.”
Reducing Human Touchpoints
Current freight logistics involves numerous human interactions for each shipment. “Every time a container moves, it gets touched by different persons,” Nilanjan explains. “There will be third parties involved, like freight forwarders, third-party logistics providers (3 PLs), or other intermediaries with human interventions.” These multiple touchpoints create opportunities for delays and mistakes. “Most communications happen over email, and there are relationships between all of them, so they know each other and love to do business together,” he adds. This relationship-based system introduces biases and inefficiencies that impact the entire supply chain.
The solution lies in modeling carriers and BCOs as intelligent agents using artificial intelligence. “The broader way of including AI agents is identifying proactively wherever the lag is,” Nilanjan shares. “If you have to book an ocean carrier, it takes four to six weeks to get a particular allocation, and that might not be known in advance.” By introducing AI agents, companies can reduce or eliminate these booking errors and improve efficiency. “If AI agents get involved, they know in advance when the booking needs to be done and eliminate manual mistakes,” he says. ” In the booking process, human coordination between the shipper and ocean carrier often leads to delays. AI agents can streamline this by making those interactions data-driven and automated. “
Enhancing Decision-Making with AI
Nilanjan’s company has implemented this approach with measurable results. “We have engaged multi-agent reinforcement learning where AI agents work on our behalf, interacting with carrier agents to handle negotiations,” he explains. These agents determine optimal booking times and calculate the probability of booking acceptance. The system particularly shines during unexpected disruptions. “Why do we use reinforcement learning? For unpredictable situations like the Red Sea crisis or sudden tariff changes where ocean rates suddenly increase,” Nilanjan notes. “The decision-making is monitored by agents that constantly adapt and learn based not just on previous records but also on the changing environment.”
One of the most significant advantages of this approach is leveling the playing field for smaller shippers. “If you’re a small or medium-level shipper, ocean carriers might not give you the same allocations. As ships get full, if you don’t ship regularly with those carriers, you won’t be on their priority list,” Nilanjan points out. AI agents tackle this disadvantage by making unbiased decisions. “AI agents will take risks considering the potential reward. They’re programmatically rewarded, so they’ll consider what’s best overall, not just based on relationships with carriers,” he explains.
As freight logistics continues to evolve, Nilanjan sees wider adoption of agent-based modeling. “Modeling behavioral patterns is an ongoing challenge—it’s not something easily captured. However, as the benefits become clearer in areas like cost savings and crisis management, both shippers and carriers are likely to adopt this approach. Ultimately, enabling AI agents to streamline and stabilize the supply chain serves everyone’s best interest.”
Follow Nilanjan Mitra Thakur on LinkedIn to explore AI-driven freight solutions.
