Trade Win No. 9
AI TRANSFORMS TRADE COMPLIANCE SERVICES, WITH HUMANS STILL IN THE LOOP | Kristina Iotzova, Kholofelo Kugler, and Esther Okoro
For decades, trade compliance was almost entirely a manual undertaking. Goods classification, origin determination, tariff calculations, and regulatory monitoring traditionally relied on teams of customs officers and trade compliance professionals reviewing regulations, spreadsheets, and supporting documentation. That landscape is changing rapidly. Digitalization, automation, and Artificial Intelligence (AI) are reshaping these services. According to a 2025 study led by the International Chamber of Commerce and the World Trade Organization (WTO), businesses are increasingly using AI for market intelligence, trade compliance, customs classification, and tariff calculations. A 2024 WTO report also states that 25% of customs administrations were already using AI and Machine Learning (ML), while a further 25% were planning to adopt them. AI and ML solutions enhance process efficiency and effectiveness, reduce trade costs, and support data-driven decision-making, improving customs and trade compliance operations for all stakeholders.
A market in transition
Evidence of AI’s growing role in trade is clear. The World Customs Organization reports that AI and ML can transform customs administrations across fraud detection, goods classification, risk assessment, and cargo clearance, signalling a structural shift in border management.
Moreover, researchers have shown that ML models can explain around 70-73% of the variation in illicit trade. An Inter-American Development Bank study found that, compared to traditional ML models, LLMs like GPT-3.5 yielded higher accuracy in product classification, achieving 60–70% accuracy at the HS 6-digit level and up to 80–90% at the 2-digit level.
Companies are deploying AI-powered solutions that automate product classification, tariff calculation, and regulatory monitoring across supply chains in real time. These solutions substantially reduce manual processing, decrease human error, uncover preferential duties, and integrate seamlessly with Enterprise Resource Planning systems.
Case Study: Besso Ltd.
Besso Ltd. is a Swiss startup illustrating how the private sector is operationalising AI in trade compliance and supply chain management. Engaging with multiple companies across import, export, manufacturing, and logistics, Besso has identified a consistent pattern: trade compliance is highly manual, fragmented, and resource intensive. Besso’s platform turns fragmented trade and supply chain data into actionable intelligence. Developed through close collaboration between ML engineers and trade specialists, the platform translates trade regulations into scalable automated workflows. In doing so, Besso enables businesses to address regulatory change proactively, identify cost-saving opportunities, and scale trade compliance activities efficiently. Besso’s platform helps compliance teams use 25% more capacity for high-value, strategic work instead of manual tasks.
Automation and human expertise go hand-in-hand
Despite their power, AI and automation cannot operate fully autonomously. Automated systems are only as good as the data and ontologies they integrate, which human experts design, validate, and refine. Without human oversight, AI-powered systems risk scaling errors and exacerbating unpredictability.
Human trade expertise remains essential to steer automation. Automation can simplify processes, but the law often resists oversimplification. Regulations contain cross-references, ambiguities, and embedded policy rationales, while jurisprudence shapes their interpretation. Applying rules to complex factual scenarios requires judgment and contextual reasoning. Beyond textual analysis, trade compliance demands accountability, stakeholder management, and trust that is intrinsically stronger in human-to-human interactions. The future of trade compliance lies in combining human expertise with AI.
