AI Adoption Helps Small Study-Abroad Agencies in China Perform Sustainably, Survey Finds

Artificial intelligence has become one of the most talked-about tools in modern business, but rigorous evidence about what it actually does for small companies has been slow to accumulate. A new study published in the journal Discover Sustainability offers a detailed look at how AI adoption shapes the performance of small and medium-sized enterprises in a sector rarely examined by researchers: overseas study agencies in China. These agencies, which help students navigate applications, visas, and admissions to foreign universities, form a dense ecosystem of service-oriented small businesses that face intense competition, shifting regulations, and volatile international student flows. The research, led by Yang Hu, Muhammad Faliq Abd Razak, and Sufian Jusoh of Universiti Kebangsaan Malaysia, together with Ahmad Rizal Bin Mohd Yusof, suggests that AI does not simply deliver benefits on its own. Instead, its influence on sustainable performance runs partly through a broader process of digital transformation, echoing a theoretical framework that has animated management science for decades.

The study set out to answer three linked questions: whether AI adoption is associated with digital transformation, whether AI adoption is associated with perceived sustainable performance, and whether digital transformation acts as a bridge between the two. Sustainable performance, in this context, refers to an organization’s perceived ability to maintain strong outcomes over the long term rather than achieving short-lived gains. The authors gathered perceptual survey data from 554 individual respondents working at sampled overseas study agency SMEs in China, asking them to characterize their organizations’ technological practices and performance. Perceptual measures of this kind capture how employees and managers judge their own firms, which the researchers treat as a meaningful indicator of organizational realities in small businesses where formal performance records may be inconsistent or incomplete.

Methodologically, the team employed a quantitative approach and tested their hypotheses using partial least squares structural equation modeling, a statistical technique widely used in business and information systems research. PLS-SEM, as it is commonly abbreviated, allows researchers to estimate relationships among multiple constructs simultaneously, including latent variables that cannot be observed directly and must instead be inferred from multiple survey indicators. It is particularly well suited to exploratory and theory-building work with moderately sized samples, and it produces path coefficients that describe the strength and direction of the associations between constructs, along with R-squared values that indicate how much of the variation in an outcome the model explains. The authors also reported bootstrapping results, a resampling procedure used to assess the statistical stability of estimated path coefficients and the significance of indirect, or mediated, effects.

The results were consistent across the hypothesized pathways. AI adoption was positively associated with digital transformation, meaning that firms whose employees reported greater use of AI tools also reported more advanced overall digitalization of their operations. AI adoption was likewise positively associated with perceived sustainable performance. Digital transformation itself was positively associated with perceived sustainable performance, and, crucially, it partially mediated the relationship between AI adoption and sustainable performance. In plain terms, part of the benefit that AI appears to confer on these small firms travels through the wider process of digital transformation: adopting AI seems to push organizations to restructure workflows, integrate systems, and modernize processes, and it is this transformation that in turn supports durable performance. The word partially matters here, because a direct pathway from AI adoption to sustainable performance remained after accounting for the mediating role of digital transformation.

The authors interpret these findings through the lens of Dynamic Capability Theory, a framework in strategic management that holds that firms survive and thrive not merely because of the resources they own but because of their capacity to sense opportunities, seize them, and reconfigure their internal structures as environments change. For small service firms operating in the volatile market for international education, that capacity is arguably everything. Application requirements shift, visa policies change with little notice, and demand for study destinations can swing dramatically from one year to the next. According to the dynamic capabilities perspective, AI adoption can be understood as a technological resource that enhances a firm’s ability to sense and respond, while digital transformation represents the deeper reconfiguration of routines and capabilities that turns new technology into lasting advantage. The empirical pattern reported in the study, with AI acting on sustainable performance both directly and indirectly through transformation, fits this theoretical account.

The choice of research setting gives the findings particular interest. Overseas study agencies in China are overwhelmingly small businesses, and they sit at the intersection of education services, international mobility, and consumer technology. Many have already experimented with AI-powered chatbots for customer inquiries, automated document screening, and data-driven matching of students to programs and institutions. Yet SMEs in general are underrepresented in the research literature on AI and sustainability, which has tended to focus on large corporations with abundant data and formal sustainability reporting. By focusing on a service-oriented SME sector in a single national context, the study provides a granular view of how the AI-to-performance relationship plays out where resources are thin and adaptation is constant. The authors note that limited research has previously examined these relationships among SMEs, particularly in the context of overseas study agencies in China, positioning the work as an early empirical contribution in this niche.

The study also carried the procedural markers of contemporary research governance. Ethical approval was granted by the Research Ethics Committee of Shaanxi Vocational and Technical College prior to data collection, under approval ID SVTC20250518003, and the authors state that the work was conducted in accordance with the approved protocol and the principles of the Declaration of Helsinki. Informed consent was obtained from all participants, participation was voluntary, and respondents were told they could withdraw at any time without penalty, with anonymity and confidentiality protected throughout. The authors declare no conflicts of interest, and the research received no external funding. The article was published open access on 14 October 2026, after being received on 2 June 2026 and accepted on 5 October 2026, and it is citable under the DOI 10.1007/s43621-026-04948-2.

As with any survey-based study, the findings come with inherent qualifications that the research design itself makes visible. The data are perceptual and cross-sectional, reflecting respondents’ views of their organizations at a point in time rather than objective longitudinal performance records, so the results describe associations among perceptions rather than proven causal effects. The sample is drawn from a single sector in a single country, which strengthens internal consistency but limits how far the conclusions can be generalized to manufacturing firms, other national contexts, or larger enterprises. Mediation analysis in structural equation modeling identifies statistical pathways consistent with theory, but establishing that digital transformation truly causes part of AI’s effect on performance would require additional designs, such as time-lagged data or experimental interventions. These constraints are standard in this research tradition, and the authors frame their contribution as providing empirical insights consistent with Dynamic Capability Theory rather than as definitive proof of mechanism.

For practitioners and policymakers, the study’s practical implications are straightforward even if its evidence is correlational. If AI adoption supports sustainable performance partly by fueling digital transformation, then simply purchasing AI tools may not be enough; small firms may need to pair technology adoption with deliberate changes to processes, skills, and organizational structure to capture the full benefit. Governments seeking to promote digitally enabled, sustainable development among SMEs could, on this reading, focus not only on subsidizing technology but on supporting the transformation capacity, such as training and consulting, that converts tools into capability. For the thousands of small agencies that shepherd Chinese students toward foreign campuses each year, the research offers a data-driven hint that the path to durable performance may run through the combination of intelligent tools and the organizational change that surrounds them.

Subject of Research: AI adoption, digital transformation, and sustainable performance in Chinese overseas study agency SMEs

Article Title: The relationship between AI adoption and sustainable performance through digital transformation among overseas study agency SMEs in China

Article References: The relationship between AI adoption and sustainable performance through digital transformation among overseas study agency SMEs in China. (n.d.). https://doi.org/10.1007/s43621-026-04948-2

Image Credits: AI Generated

DOI: 10.1007/s43621-026-04948-2

Keywords: artificial intelligence, digital transformation, sustainable performance, SMEs, overseas study agencies, China, Dynamic Capability Theory, PLS-SEM, survey research, sustainable development, service sector, relationship