AI is often discussed through models, chips, and data centres. For most small and medium-sized businesses, however, the closer question is whether AI can reduce the cost of finding customers, understanding market requirements, communicating with partners, and handling documentation. New findings from the World Trade Organization (WTO) and International Chamber of Commerce (ICC) show that this is no longer theoretical: firms using AI in trade report benefits, yet gaps in skills, data, and regulation still determine who can turn a tool into competitive capability.
The WTO–ICC survey, released in December 2025, gathered responses from 158 businesses across regions and sectors. In this group, 63% were micro, small, and medium-sized enterprises; almost half reported using AI, and 79% engaged in international trade. Among firms already using AI, close to 90% reported benefits in trade-related activities. These figures do not represent every business or prove causation, because the sample is limited and self-reported. They nevertheless provide a notable market signal: AI is entering specific operating stages of trade, not remaining only in large technology projects.
Value begins with friction costs
For a smaller business, the barrier to selling beyond a familiar market is often not only the product. It is also the time needed to find information, read partner requirements, manage language differences, prepare documents, follow regulations, and respond to customers across time zones. The WTO’s World Trade Report 2025 notes that firms expect AI to reduce trade costs, especially in logistics, compliance, and communication. Small and medium-sized firms in the survey were generally more optimistic than large firms about expected savings, perhaps because these activities consume a larger share of their resources.
The key is not to read “AI reduces costs” as an automatic promise. A translation tool can help a sales team respond faster, but it does not replace checking contract terms or understanding a market’s practices. A regulation-summarization tool can reduce search time, but it cannot itself confirm that a shipment fully complies. Value appears when the tool sits in a workflow with clear inputs, an appropriate reviewer, and evidence that can be retained.
| Trade activity | How AI can help | Management control to retain |
|---|---|---|
| Market research | Consolidate demand, competitor, and keyword signals by market. | Check sources, data coverage, and assumptions before making a decision. |
| Customer communication | Translate, draft replies, classify requests, and track recurring issues. | The owner approves commitments, pricing, and sensitive content. |
| Documentation and compliance | Extract information, compare lists, and identify missing fields. | A specialist confirms product codes, terms, and market requirements. |
| Logistics operations | Summarize exceptions, anticipate information needs, and prioritize work. | Operations remains responsible for data, schedules, and delivery commitments. |
The real gap is data and skills
The WTO–ICC findings also reveal a notable gap: 66% of businesses in high-income economies in the survey said they had adopted AI, compared with 27% in low- and lower-middle-income economies. Firms identified familiar obstacles: regulatory uncertainty, inadequate high-quality data, and limited skills. For a Vietnamese business, the message is not to wait until every condition is perfect. It is a reminder that buying a tool does not itself create reliable data, a shared workflow, or someone accountable for the output.
A better starting point is one high-frequency trade situation with manageable control risk. A furniture exporter, for example, can use AI to consolidate feedback from overseas customers, sort requests by product, deadline, and quality issue, and draft a response in an appropriate language. The responsible employee still checks delivery terms, fulfilment capability, and every commitment. After several weeks, the company can measure response time, rework, and recurring causes rather than merely asking whether people find AI “useful.”
Example: preparing an offer for a new market
A food company wants to approach distributors in a new market. The team can use AI to structure public information on customer segments, sales channels, and labelling requirements; compare a draft offer against a partner’s questionnaire; and prepare a list of points to verify. The team should not, however, let AI conclude that a product meets regulations or send a quotation automatically. The export owner needs to verify the sources, commercial terms, supply capability, and final version. AI reduces the work of finding and organizing information; market decisions still require evidence and company accountability.
From trying tools to building trade capability
Three principles can help a business avoid fragmented experimentation. First, begin with a friction point in a process rather than with a tool feature. Ask: which task is time-consuming, repetitive, and has an output clear enough to check? Second, identify which data may be used and who reviews output before it leaves the organization. Customer, price, contract, and shipment data should not enter a process merely because doing so is convenient.
Third, measure value together with risk. A response-time or throughput measure is necessary, but not enough. The business should also track errors, rework, exceptions, and customer feedback. When results are good, the process can be standardized, taught, and selectively scaled. When errors appear, the organization knows whether it needs to adjust data, instructions, access rights, or a review point.
Conclusion
AI can make international trade more accessible by reducing friction costs in information, language, and administration. The benefit does not come from replacing people with a chatbot. It comes from designing processes in which AI helps find, organize, and draft, while people check, decide, and remain accountable. For small and medium-sized businesses, that is the practical route from a new tool to a capability that serves the market.
References
World Trade Organization. (2025). World trade report 2025: Making trade and AI work together to the benefit of all. https://www.wto.org/english/res_e/publications_e/wtr25_e.htm
World Trade Organization. (2025, December 11). WTO-ICC business survey sheds light on opportunities and challenges for AI use in trade. https://www.wto.org/english/news_e/news25_e/ecom_11dec25_240_e.htm
