AI Agents Deployed for Logistics and Telecom

A "rolled container" occurs when cargo isn't loaded onto its scheduled vessel, often due to carrier overbooking, port congestion, customs holds, or documentation errors. This disruption creates shipment delays and adds complexity for supply chain managers, forcing them to find new bookings. Project44's AI agent operates on its Movement decision intelligence platform, which processes data from a network of over 259,000 carriers and 1.5 billion annual shipments. The agent autonomously monitors for roll risk, confirms the issue with carriers, and assembles alternative voyage options for a human analyst to approve. Early results show the AI can identify roll risk up to 35 hours before official carrier status updates. This reduces the time needed to detect the problem and prepare a rebooking recommendation from several hours of manual work to under five minutes. In telecommunications, generic "frontier" AI models often fail because they struggle to interpret complex network data and automate operations with the required accuracy. Data from GSMA Intelligence indicates that only 16% of generative AI deployments in the telecom sector are used for network operations. The GSMA's Open Telco AI initiative aims to solve this by creating a shared ecosystem with telco-specific AI models, specialized datasets, and benchmarking tools. The goal is to foster open collaboration between operators, vendors, and AI developers to build AI that understands the unique language and standards of telecommunications. Founding partners for the initiative include AT&T, which is releasing a family of open-source models trained on public data, and AMD, which is providing the GPU computing power for model training and evaluation. Other contributing partners include Nvidia, Orange, SK Telecom, Telefónica and Vodafone.