AI’s real cargo test in Europe - Air Cargo Week
AI is being positioned as a decision-support tool rather than a replacement for cargo professionals, helping connect fragmented data, reduce manual work and accelerate decisions across pricing, quoting,. . .
AI is being positioned as a decision-support tool rather than a replacement for cargo professionals, helping connect fragmented data, reduce manual work and accelerate decisions across pricing, quoting, booking, tracking and planning. The biggest near-term gains are expected from embedding AI into existing workflows, automating repetitive tasks such as email processing, rate searches, capacity comparisons and shipment tracking while ensuring decisions remain connected to real operational constraints. The next phase of AI adoption is moving towards intelligent freight networks and agentic systems that can anticipate disruptions, recommend actions and improve collaboration across airlines, forwarders, handlers and shippers, with success measured by faster and better decision-making. Rather than replacing cargo professionals or creating fully autonomous logistics networks, technology providers see AI’s strongest near-term role in reducing manual work, connecting fragmented data and helping teams make faster decisions across pricing, quoting, booking, tracking and planning. “The future of air cargo isn’t autonomous logistics. It’s real-time, intelligent freight networks where AI turns market signals into decisions before humans would even know they needed to act,” Ian Arroyo, Chief Strategy Officer at Freightos, said. That distinction matters in an industry where many commercial and operational decisions still depend on emails, spreadsheets, partner portals and disconnected systems. Executives from Accelya, CargoAi and Freightos pointed to the same central challenge: AI can only deliver value if it is embedded into real cargo workflows and connected to reliable data. For Alessandro Alfano, SVP, Product & Technology at Accelya, the value of AI is not in adding another standalone digital layer, but in joining the decisions that already shape airline cargo operations. “At Accelya, we see AI less as a standalone tool and more as a layer of intelligence running across the airline workflow,” he said, pointing to the need to connect offer, booking, capacity, execution, delivery and settlement. In cargo, Alfano said, AI can be applied to capacity forecasting, dynamic pricing, offer creation, load planning, control tower visibility and disruption management. But he stressed that the real test is whether a decision can actually be executed once it has been made. “Intelligence is only valuable when it is executable, when price, capacity, route, handling constraints and settlement path all line up before the commitment is made,” he said. Air cargo tensionsThat reflects one of the core tensions in air cargo digitalisation. Faster responses are only useful if the shipment can still move as promised. A quote generated quickly has limited value if capacity, restrictions, routing or handling requirements have not been properly checked. Matt Petot, CEO of CargoAi, made a similar point, arguing that AI becomes useful when it is built directly into daily operational processes. “At CargoAi, we believe AI becomes truly valuable when it is embedded directly into operational workflows rather than existing as a standalone tool,” he said. CargoAi says it has spent the past seven years building an AI layer across its CargoMART ecosystem. Its CargoCoPilot AI Agent enables freight forwarders and airline teams to search rates, compare flight options, generate quotations, track shipments and access operational information using natural language across CargoMART, email, WhatsApp and APIs. The company has also developed the CargoMART Outlook Email Plugin, which it says automatically turns unstructured customer emails into structured shipment requests. That allows users to move from email to quotation or booking without manually re-entering shipment details. According to CargoAi, customer emails can be transformed into structured shipment data and used to initiate searches, quotes or bookings faster, reducing repetitive workload by up to 50 percent. That figure should be treated as a company claim, but it points to one of the clearest areas where AI is being applied: the removal of repetitive work from high-volume cargo processes. “The greatest productivity gains are happening in areas where manual work, emails and fragmented systems still dominate,” Petot said, listing quotation management, rate search, capacity comparison, booking creation, shipment tracking and customer service as examples. Arroyo also identified the strongest gains in areas combining high transaction volume and structured data, including pricing, quoting, booking, documentation, customer service and exception management. “What’s interesting is that the most valuable use cases aren’t necessarily the most exciting ones,” he said. “The boring stuff is where the value is. Companies often look for transformational AI use cases when they should be looking for friction. Remove enough friction and the transformation takes care of itself. ” That practical focus runs through the current AI discussion in air cargo. The opportunity is not only to automate tasks, but to shorten the distance between a signal and a decision. Arroyo said freight has historically been managed through periodic decision-making, with procurement taking place annually, rates updated monthly and capacity planning done quarterly. In his view, the market no longer behaves that way. “The challenge isn’t getting more information. The challenge is knowing what matters and acting on it quickly,” he said. For airlines, that has implications for pricing and capacity management. Alfano said AI can help airlines move away from fixed rate cards or static rate sheets towards pricing decisions that reflect live market signals, customer context and the capacity available at the booking moment. He also highlighted interline coordination as a use case. In recent industry work with IATA, he said AI agents supported multi-airline interline booking scenarios, while deterministic routing logic tested proposed options against live capacity, routing and operational constraints before they moved closer to confirmation. The result, according to Alfano, was a move away from manual, email-based coordination towards a more structured and auditable process. “The objective is simple: to move cargo away from reactive, email-led co-ordination and towards decisions that are connected, system-driven and grounded in operational reality,” he said. Tracking and disruption management are another area where AI is being positioned as a practical tool. CargoAi’s AI Predictive Tracking & Alerts solution uses machine learning models trained on millions of historical shipment events, enriched with live airline data, to forecast cargo milestones. Petot said traditional tracking tells teams what has already happened, while predictive tracking helps teams anticipate what may happen next and act before a disruption affects the customer. Planning shiftAccelya’s Alfano made a related point, saying that if delays, cancellations, offloads or data mismatches are identified earlier and managed through structured workflows, disruption becomes less of an improvisation and more of a controlled response. For Freightos, the more important planning shift is not simply better forecasting, but faster adaptation. “I think the industry is focusing on the wrong word. The breakthrough isn’t forecasting, it’s adaptation,” Arroyo said. He argued that the last five years have shown how fragile forecasts can be in the face of tariffs, geopolitical events, pandemics, weather disruptions and conflict. The companies that perform best, he said, may not be those that predict the future most accurately, but those that respond fastest when conditions change. “AI allows planners to model multiple scenarios simultaneously and understand tradeoffs in minutes rather than days,” he said. That view is echoed in the way CargoAi describes its CargoMART Strategic Hub and Live Rates tools, which it says provide real-time views of market competitiveness, allocation performance, quality and carrier benchmarking. Petot said this helps airlines and forwarders make decisions based on what is happening now, rather than what happened weeks ago. The labour question is more nuanced. None of the executives interviewed presented AI as a substitute for cargo expertise. Instead, they framed it as a way to support stretched teams, reduce repetitive work and make operational knowledge more accessible. “AI should not be cast as a substitute for people,” Alfano said. “Air cargo is too complex for that, and too dependent on experience, judgement and local knowledge. ” Petot said AI can help remove repetitive manual work, accelerate access to information and support users with recommendations inside their daily workflows. That could allow experienced employees to focus on higher-value decisions, while newer staff become productive faster. Arroyo described the same shift as scaling expertise. “Air cargo remains one of the most knowledge-intensive industries in the world,” he said. “The challenge is that much of that knowledge still sits inside the heads of experienced operators. AI allows organisations to capture, distribute, and operationalize portions of that expertise. ” Freightos framed the issue in similar terms. “Start with a business problem. Not an AI project,” Arroyo said. “If you can’t connect AI to a business metric, you’re probably running an experiment rather than creating value. ”Agentic AI systemsLooking ahead, the three companies see AI moving beyond analysis into more active support for cargo workflows. Accelya points to agentic AI systems that can identify opportunities, recommend actions and, within defined approval boundaries, trigger next steps. CargoAi says its MCP Connector extends CargoMART capabilities into AI assistants such as ChatGPT, Claude and Microsoft Copilot, allowing users to search rates, compare flights, retrieve airline contacts, track AWBs and access operational intelligence using natural language. Petot said the next generation of air cargo will be “AI-native”, with AI agents supporting customer tracking replies, best-rate searches, critical AWB monitoring, airline escalation drafts and GHA lookups. For Arroyo, this points towards what he calls Intelligent Freight Networks, where data from carriers, forwarders, ground handlers, customs brokers and shippers is connected and turned into actionable decisions in real time. “The companies that win won’t be the ones with the most AI,” he said. “They’ll be the ones that make the best decisions the fastest. ”