More AI is not Always Better For Small Firms Chasing Export Growth

Artificial intelligence (AI) is becoming a practical trade tool for small exporters that cannot afford large technology teams, overseas offices or dedicated market research units. A new study of small and micro-enterprises in China's Yiwu International Trade City finds that AI can improve export performance, but it also warns that more technology does not automatically mean better results.

The study, titled "Artificial Intelligence and Export Performance in Small and Micro-Enterprises: The Roles of Internal Capability and External Tools," was authored by Mengyang Gu and Chuyue Jin of Kookmin University in the Republic of Korea and published in Sustainability. Based on survey data from 475 exporting small and micro-enterprises, it examines how two AI-related resources shape export outcomes: ready-made external AI tools and firms' own internal AI capability.

The findings move the AI discussion beyond adoption hype. Both external AI tools and internal AI capability are linked to stronger export performance. According to the study, when firms invest heavily in both, the added benefit of each may decline. For resource-constrained small firms, overlapping AI investments can duplicate functions, stretch managerial attention and reduce the value gained from scarce capital.

The new trade assistant for small exporters

For small exporters, AI is no longer an abstract technology used only by large companies. AI tools can translate buyer messages, generate product descriptions, draft marketing content, support online promotion, answer customer inquiries and search for overseas market information.

Small and micro-enterprises often face the same global competition as larger firms but with fewer resources. Language barriers, limited staff, weak market intelligence and slow customer communication can hold them back in export markets. AI tools can reduce some of those barriers by making routine cross-border tasks faster and cheaper.

Known as a major small-commodity export hub, Yiwu is home to thousands of small trading businesses serving global buyers. Digital trade platforms such as Chinagoods have introduced AI-enabled services for translation, content generation, customer communication, digital marketing and trade matching. These are not futuristic experiments, but practical tools used in daily export operations.

The study separates AI use into two categories:

• External AI tool utilization refers to platform-based or service-provider tools that firms can use without building technology in-house. These include AI translation, content generation, automated customer communication and digital marketing support.
• Internal AI capability refers to a firm's own ability to select, integrate, manage and embed AI into business routines through skills, data readiness, processes and managerial support.

Export performance was measured through firms' perceived export outcomes, including export sales ratio, export sales growth, export profit contribution and overall export performance. The researchers used regression analysis to test the individual and combined effects of external AI tools and internal AI capability.

Ready-made AI tools open the export door

The first major finding is that firms that used external AI tools more actively reported better export performance. In practical terms, a small exporter that can translate buyer messages quickly, generate multilingual product descriptions and respond to customer inquiries faster has a stronger chance of converting interest into sales.

External AI tools are especially valuable because they lower the cost of digital participation. A small firm does not need to develop an algorithm, hire a data science team or build complex infrastructure. It can access standardized tools through platforms and apply them to daily export tasks.

In emerging and developing economies, many small businesses struggle to cross borders because they lack language skills, market information and digital marketing capacity. Affordable AI tools can help level part of that field, allowing smaller firms to reach overseas buyers more efficiently.

Another key finding is that internal AI capability also improves export performance, which means firms benefit not only from using AI tools, but from knowing how to integrate them into actual business processes. A firm with stronger internal AI capability is better able to decide which tools are useful, train staff, manage data, redesign workflows and embed AI into customer service, marketing and export coordination. Without that internal discipline, AI tools may remain occasional add-ons rather than productivity-enhancing resources.

This is crucial for policymakers and development agencies. Digital access alone is not enough. Giving small firms tools without helping them build the ability to use those tools effectively may produce shallow adoption. Training, managerial awareness and process redesign matter as much as software access.

The hidden cost of doing the same AI job twice

According to the analysis, external AI tools and internal AI capability do not always reinforce each other. Their interaction is negative and significant, meaning the benefit of one becomes smaller when the other is already highly developed.

Does this mean that AI investment is harmful? Not at all. Both forms of AI resource are positively linked to export performance. The issue is duplication. In Yiwu's export environment, external tools and internal capability often support the same routine tasks, such as translation, product content generation, buyer communication and market information processing.

When a small firm invests heavily in both, it may end up paying twice for similar functions. It may also create new coordination burdens: staff must learn multiple systems, managers must monitor more tools and scarce attention gets pulled away from other export priorities such as product quality, logistics, customer acquisition or relationship building.

The study describes this as partial substitution. Once one AI resource is strong enough to handle a task, adding another resource for the same function may bring smaller returns. The finding is especially important for small and micro-enterprises because their resources are limited. A large firm can absorb some duplication but a small exporter often cannot.

Policy should fund judgment, not just tools

Small firms are often encouraged to digitalize quickly, but digitalization can become costly if it is not strategic. Many micro-enterprises operate on thin margins. A poorly planned AI investment can drain time and money without improving competitiveness.

The research suggests that export support policies should shift from "adopt more AI" to "adopt the right AI in the right place." Firms with weak internal systems may gain more from accessible external tools. Firms with stronger digital routines may need more customized internal integration rather than heavier reliance on standardized platform tools.

When it comes to inclusive trade, AI can help small firms participate in global markets, but better-resourced firms may capture more benefits if they have stronger skills and managerial capacity. Without targeted support, AI could widen the gap between digitally prepared enterprises and those adopting tools without strategy.

Small exporters need support that helps them make better technology choices, not just subsidies for more tools. Export agencies can help firms identify which tasks AI should handle first: translation, buyer communication, product listings, marketing content or market research. Business support programmes can teach firms how to compare external tools with internal capability needs, avoid duplication and measure whether AI use improves export outcomes.

Development agencies and international organisations can treat AI capability as part of trade capacity-building. Support could include digital skills training, shared AI services, advisory programmes, low-cost platform access and guidance on data use. The goal should be practical productivity, not technology adoption for its own sake.

Risks, limits and unanswered questions

The study focuses on a real export cluster where AI tools are already part of daily business. It also makes a useful distinction between external tools and internal capability, which gives the findings practical relevance. However, the study is concentrated in Yiwu, a highly platformized export hub, so the same pattern may not apply equally in less digitalized regions or other countries. The data come from self-reported surveys, and export performance is based on perceived outcomes rather than verified transaction or financial records. The study is also cross-sectional, so it cannot fully prove causality.

Future research should test similar questions in other export clusters, industries and countries. Longitudinal studies could show whether using external AI tools helps firms build internal capability over time or discourages them from doing so. Objective export data from platforms, customs records or financial statements would also strengthen the evidence.