SPECIAL ADVERTISING SECTION
AI IN LOGISTICS
enforceable and not stable. We help our customers deal with volatility and risk.”
Enhancing performance, decisionmaking
By Raft’ s estimates, transporting an international shipment— from China to the United States, for example— involves 20 or more parties, starting with the manufacturer.
Key customer benefits are efficiency, capacity creation and error reduction, Mehta noted. Raft’ s customs, accounts payable and customer bookings functions see the most requests from clients.
“ We’ ve seen customers who have been able to process double the number of units of work with the same team, while others have reduced the size of their document processing teams by 70 % to 75 %,” he said.“ Errors have fallen 80 % to 90 %.
“ It varies from customer to customer and use case to use case, but we’ re seeing multiples like 2x, 3x, 4x, 5x improvements in capacity for complex use cases.”
Where AI is concerned, Xeneta’ s focus has been more on traditional models, although generative AI is becoming increasingly prominent. Machine learning is integral to its data ingestion and analysis processes, underpinning everything from rate validation to outlier detection.
“ We get rate sheets from many different customers in many different formats working with many different suppliers, and each has their own different Excel file that they use,” Brocca said.
AI enables the company to aggregate, ingest and anonymize data to“ figure out if the rate is correct” he explained. AI is also integral to outlier detection and validation.
Xeneta’ s three- and six-month forecasting is based on machine learning and goes beyond giving price predictions.
“ Explainability is very important in this industry. If I just give you a number as to what the rate will be six months from now, that’ s not very actionable,” Brocca said.“ I need to be able to explain why we think that and what factors affect the model. So, we’ ve found a successful way to work with our market analyst team and keep the human in the loop to provide context and explain why we make a certain prediction.”
Xeneta uses generative AI in its customer-facing services to provide insights and recommendations.
“ You can see tons of different data sets but making sense of these and figuring out what to do next requires a lot of effort. It’ s impossible to do at scale across all your lanes, across all your contracts for your suppliers. So, we use AI to simplify that [ through ] our AI agent,” Brocca said.
“ You select what output you are looking for. For example, you tell us you want to optimize your supplier mix or optimize your spend— then we’ ll have [ AI ] agents running in the background collecting data for you, looking at your rates, comparing with the market, looking how the market is shifting.
“ We train the model with a lot of our insights and data and market expertise. Essentially the goal is that you can wake up in the morning and we’ ll tell you‘ Here are the key three things you need to do today.’”
Environmental, regulatory considerations
Artificial intelligence requires tremendous computer processing power and therefore energy. A typical AI-focused data center consumes as much as electricity as 100,000 homes, while the largest under-construction facilities will use 20 times as much, according to the International Energy Agency( IEA).
Practical AI for forwarders and brokers
Global logistics doesn’ t need transformation projects. It needs wins you can measure next week: 15 % less rework, two-day faster cycle times, zero errors when volumes spike by 40 %.
Start with one pain point. Pick the process where mistakes cost you the most— customs documentation errors, shipment exceptions eating up your best employees’ time, or quote turnaround killing deal velocity. Deploy AI there first. Measure impact in days, not months.
With margin pressure intensifying and trade volumes volatile, freight forwarders and customs brokers can’ t scale headcount linearly. The question isn’ t whether to use AI— it’ s where to start for fastest ROI.
The pattern that works Target one high-volume, high-error process. Automate repetitive classification, data entry, or document review. Let your people handle exceptions and client relationships. Measure time saved and errors eliminated. Then expand.
Results show up fast: Customs preparation time falls from 60 minutes to under 10, with clearance 30 % to 40 % faster; AP automation reaches 40 % zero-touch validation; exception management detects issues earlier, reducing handoffs and strengthening on-time performance.
Proven at scale
Imperative Logistics consolidated 16 systems to achieve 90 % faster document turnaround, cut errors from 20 % to under 10 %, and reduced manual touch to less than 1 %— unlocking 2-3x capacity with the same team. Some workflows now run in 5 minutes instead of 3 hours.
Orchestration across systems
AI only delivers when your systems talk to each other. Raft orchestrates across your TMS, ERP, customs platforms and finance systems via API, syncing data so AI works on complete, accurate information.
Raft is trusted by 60 + forwarders and customs brokers, including GEODIS, RIM Logistics and Logistics Plus, processing 5 + billion data points annually. See how Raft orchestrates customs and freight operations: raft. ai.
48 Journal of Commerce | December 1, 2025 www. joc. com